Public evidence summary
Test-status register updated 30 September 2026
- Record date
- 30 September 2026 snapshot; individual historical test dates are shown.
- Scope
- Published register or policy snapshot; per-item scope and dates apply.
- Current deployment
- Not independently verified by this historical evidence. Current availability
Recorded result
The 30 September 2026 register contains 98 test records: 96 marked green, one failed and one not requiring a rerun. It adds the dated agent-session production canary. The failed encrypted soak remains recorded even though its stated later rerun succeeded; a not-required entry remains distinct from a passing rerun. Each row retains its historical test date and scope. This updated register is not a new execution of all 98 tests, a universal agent-compatibility result, or independent verification of today’s deployment.
Individual recorded test outcomes
Historical opt-in encrypted-record test
green
On 7 September 2026, an engine test on hosted hardware verified that an opted-in session produced an encrypted transcript readable with the customer’s key, while an opted-out session produced no transcript in the record subsystem. Alteration of a recorded entry was detected. This record explicitly says the feature was not enabled in production. The separate, bounded response-replay cache is outside the transcript subsystem, so the opt-out result is not a claim that no response bytes can exist anywhere. Production enablement requires separately reviewed privacy disclosures and evidence.
Date: 7 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Not enabled in production in this record Replay cache outside recording subsystem Engine test, not current deployment proof
Read the dated evidence summary →Historical transcript-conversion test
green
On 7 September 2026, deterministic checks converted a real Claude Code transcript into inputs for other agent tools and produced a separate task handoff. The record reports the conversation information lost in each conversion. No live target agent was run in this test, and no successful task continuation is claimed. Each supported source/target pair needs its own live completion test before it can be advertised as harness switching. The task handoff remained usable independently of transcript conversion.
Date: 7 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Mechanics only; no live target agent Only one source-transcript reader tested Conversation loss declared per target No harness-switch availability claim
Read the dated evidence summary →Historical signup-credit and free-session test
green
On 1 September 2026, laboratory tests with real virtual-machine sessions verified automatic account activation, a one-time signup credit, concurrency limits and checkpoint-and-park behavior. Repeat signup by the same authenticated identity did not receive another grant. The record’s parking behavior was based on elapsed runtime, not inactivity, and some abuse controls were approximations. Production self-service was disabled in this historical version. A 17 September amendment changed the credit-exhaustion check to an isolated internal test; that amendment is not itself a new full production run. This record does not establish current signup grants or availability.
Date: 1 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Runtime rather than idle detection Abuse-control approximations Historical production self-service disabled Amended exhaustion test isolated
Read the dated evidence summary →Historical staff-console control tests
green
On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.
Date: 1 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.
Read the dated evidence summary →Historical staff-console control tests
green
On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.
Date: 1 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.
Read the dated evidence summary →Historical staff-console control tests
green
On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.
Date: 1 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.
Read the dated evidence summary →Historical staff-console control tests
green
On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.
Date: 1 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.
Read the dated evidence summary →Historical hosted attach-and-fork test
green
On 3 September 2026, laboratory checks created owned fork branches, exercised terminal attachment and refused another account’s access. File-divergence checking was skipped because most branch networks were unreachable, and attachment used the stable parent. A separate production narrative reports successful owned forks and keyboard interaction through the hosted client after deployment. Final robustness changes could not be rerun on the damaged laboratory environment. The laboratory assertions and production narrative cover different scopes; neither resolves the recorded branch-network limitation.
Date: 3 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: File-divergence check skipped in laboratory Attachment tested on stable parent Separate production narrative Final robustness changes not rerun on damaged laboratory environment
Read the dated evidence summary →Historical independently verified branch-work test
green
On 7 September 2026, a hosted-hardware test forked one task workspace into three branches. Two branches reached authorized model providers, performed work in their own workspaces and produced accepted candidates under independent verification. The unreachable branch was counted as failed rather than omitted. Exactly one verified result received scoped publication authorization. This was authorization for a review branch or pull request, not an actual external publication. Larger branch counts were not tested.
Date: 7 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Three branches only Unreachable attempt counted failed Authorization intent, no external publication
Read the dated evidence summary →Historical deployment-platform portability test
green
The record resolved on 30 August 2026 reports that the existing readiness, memory-continuity, lifecycle and checkpoint tests passed on the hosted x86 platform. Their test definitions were checked for consistency with the laboratory versions. The first run failed because its corruption test did not actually change the selected data; after that test was corrected, the negative test passed. This demonstrates the recorded platform transition, not continuous availability, portability across arbitrary architectures or a claim that the first failed result never occurred.
Date: 30 August 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Initial corruption test invalid, corrected in later run Named platform only No arbitrary cross-architecture portability
Read the dated evidence summary →Historical account-isolation test
green
On 30 August 2026, a two-account test on hosted hardware verified that each owner could operate its own session and that forged or cross-account attempts to read, attach, checkpoint or resume were refused without changing the target session. The test exercised the engine’s account-ownership boundary using test-generated account assertions. It was not a full end-to-end test of external sign-in or today’s document portal. Legacy laboratory sessions had a separate local-access scope, explicitly outside the owned-session claim.
Date: 30 August 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Two-account engine test Not external sign-in or document authorization Legacy laboratory-session scope differs
Read the dated evidence summary →Historical owned-session broker test
green
On 30 August 2026, a test on hosted hardware used an account token to create, read and checkpoint an owned session through the control plane. Another account’s token and a revoked token were refused. The tested components were colocated; public-domain transport, real Google/GitHub sign-in and real provider-key use were not established by this test and awaited separate setup. The result demonstrates that recorded ownership and token-revocation path, not a complete customer onboarding or payment journey.
Date: 30 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Colocated components Real external sign-in and BYOK outside scope Not complete paid onboarding
Read the dated evidence summary →Historical private service-boundary test
green
The record sealed on 31 August 2026 reports checks run on 30 August showing that the authorized control plane could create and read an owned session across the private network. Public and unauthenticated transport attempts were refused, including plaintext and untrusted client certificates. Network restrictions and certificate authentication were tested independently. This was a same-region control-path test. It did not test moving a checkpoint to another host, recovery after host loss or today’s network configuration.
Date: 31 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Run date differs from seal date Same region control path Not cross-host restore Historical configuration only
Read the dated evidence summary →Provider-key isolation and safe refusal without a key
green
On 30 August 2026, a hosted-service test used a fixture credential and a simulated provider. The session completed its request using the account’s supplied key, the inspected logs and stored records contained no plaintext copy, and a session without a usable key was refused. A negative control confirmed refusal when the key could not be decrypted. Assertions were appended on 6 September 2026. This record did not make a live provider call and does not establish that arbitrary credentials can never leak.
Date: 30 August 2026; 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Simulated provider and fixture credential Live provider behavior belongs to a separate record No general proof against credential leakage
Read the dated evidence summary →Historical hosted-client and provider-key trial
green
The record sealed on 31 August 2026 reports hosted-client checks run on 30 August: a token created a session on the live fleet, read its usage information and stopped it. A revoked token was refused before creating a session. A separate team-operated trial on 31 August completed a real model call using the team’s own provider key. That real-key trial is narrative evidence, not one of the automated gate assertions. These two scopes are distinct and do not establish behavior for every workload or current release.
Date: 31 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Automated client checks versus narrative real-key trial Team-owned key, not customer trial Historical availability only
Read the dated evidence summary →Historical restart-and-reboot recovery test
green
On 31 August 2026, a hosted-hardware test verified that the services restarted after a forced process stop and returned automatically after the host rebooted. A previously checkpointed and stopped counter session resumed with its in-memory witness preserved and its counter advancing. The negative test detected the absence of automatic restart. This test used a small counter specimen and a saved checkpoint; it does not promise preservation of unsaved work, arbitrary workload recovery or failover to another host.
Date: 31 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Counter specimen Checkpointed and stopped before reboot Not cross-host failover or unsaved-work guarantee
Read the dated evidence summary →Historical erasure-failure isolation regression
green
A regression run on 7 September 2026, sealed in the 8 September record, verified that a session with destroyed account key material could not resume while another account’s session resumed immediately afterward without restarting the service. The earlier reported service failure was traced to an invalid test request, not damage to the service. This was an isolated engine test on hosted hardware. It did not exercise the production account-deletion route, which requires its separate canary evidence.
Date: 7 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Earlier failure was invalid test request Isolated engine test Production deletion route separate
Read the dated evidence summary →Historical account-deletion cleanup correction
green
On 8 September 2026, regression and production-canary evidence reported that deleting a synthetic account removed its active-session resources before destroying its key. The prior behavior left unusable session residue and caused health checks to fail. The corrected deletion refused to report success if teardown failed. Billing-related tombstone records were retained under the stated policy. The production check used a synthetic tenant; unrelated resources from a concurrent test were cleaned separately. This is dated evidence for that cleanup path and does not remove later open account-deletion findings.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenant Unrelated concurrent-test cleanup separate Billing tombstones retained Later deletion findings remain public
Read the dated evidence summary →Historical account-deletion cleanup correction
green
On 8 September 2026, regression and production-canary evidence reported that deleting a synthetic account removed its active-session resources before destroying its key. The prior behavior left unusable session residue and caused health checks to fail. The corrected deletion refused to report success if teardown failed. Billing-related tombstone records were retained under the stated policy. The production check used a synthetic tenant; unrelated resources from a concurrent test were cleaned separately. This is dated evidence for that cleanup path and does not remove later open account-deletion findings.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenant Unrelated concurrent-test cleanup separate Billing tombstones retained Later deletion findings remain public
Read the dated evidence summary →Historical checkpoint-timeout correction
green
The 8 September 2026 record reports a correction for checkpoints that exceeded the previous fixed deadline. Tests verified an adjusted time allowance, bounded recovery after a timeout and explicit recording of failed checkpoints. The page reports deployment verification before the later soak resumed. Measurements came from one host and one storage class; the calibration did not track storage that degraded while the service was running. A prematurely started soak on the previous build was stopped, so it is not evidence for the corrected build.
Date: 8 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: One host/storage class Calibration not continuously tracking degradation Old-build soak stopped, not proof of fix
Read the dated evidence summary →Historical production credit-authorization correction
green
On 17 September 2026, laboratory tests reproduced and then rejected unauthorized customer credit changes after the correction. A separate production check with a disposable unfunded account confirmed that customer credentials could not change credit, balances stayed unchanged and legitimate ledger reading still worked. A point-in-time ledger review found no unexpected grant-reference categories in the surviving records. That review is not a guarantee that no misuse ever occurred. Some broader laboratory reruns were deferred under a documented exception; the scope of this production check remains credit authorization.
Date: 17 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Disposable unfunded tenant; verification granted nothing Point-in-time ledger review only Broader rerun exception Historical, not current exploitability proof
Read the dated evidence summary →Historical production credit-authorization correction
green
On 17 September 2026, laboratory tests reproduced and then rejected unauthorized customer credit changes after the correction. A separate production check with a disposable unfunded account confirmed that customer credentials could not change credit, balances stayed unchanged and legitimate ledger reading still worked. A point-in-time ledger review found no unexpected grant-reference categories in the surviving records. That review is not a guarantee that no misuse ever occurred. Some broader laboratory reruns were deferred under a documented exception; the scope of this production check remains credit authorization.
Date: 17 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Disposable unfunded tenant; verification granted nothing Point-in-time ledger review only Broader rerun exception Historical, not current exploitability proof
Read the dated evidence summary →Historical prepaid-usage settlement correction
green
On 17 September 2026, laboratory tests reproduced a defect where session usage did not consume prepaid credit, then recorded correct one-time settlement, charges only for additional usage, parking at exhaustion and resume after a top-up. Production evidence later that day recorded outstanding usage debited once and the funds restriction triggered. No affected session was running, so production parking at exhaustion remained untested. An amendment on 18 September corrected the monitoring count without changing the settlement verdict. The record does not disclose account-linked balances or prove current deployment behavior.
Date: 17 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Production running-session parking remained untested Original monitoring count corrected on 18 September Laboratory parking used a simulated service
Read the dated evidence summary →Historical prepaid-usage settlement correction
green
On 17 September 2026, laboratory tests reproduced a defect where session usage did not consume prepaid credit, then recorded correct one-time settlement, charges only for additional usage, parking at exhaustion and resume after a top-up. Production evidence later that day recorded outstanding usage debited once and the funds restriction triggered. No affected session was running, so production parking at exhaustion remained untested. An amendment on 18 September corrected the monitoring count without changing the settlement verdict. The record does not disclose account-linked balances or prove current deployment behavior.
Date: 17 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Production running-session parking remained untested Original monitoring count corrected on 18 September Laboratory parking used a simulated service
Read the dated evidence summary →Historical signup-credit unit correction
green
On 18 September 2026, isolated tests reproduced a wrong-unit configuration that delivered less signup credit than advertised, then passed after the correction. The new checks rejected the mistaken unit and used the approved price-book value when encountering the stored error. The production section reports deployment and remediation of the undercredited account; repeating the remediation did not add credit again. The automated assertions used an isolated test identity provider with no live fleet or model calls. Production remediation is a separate recorded observation, not an independently repeated audit test.
Date: 18 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Isolated automated test Production remediation separately reported No live fleet/model calls in gate
Read the dated evidence summary →Historical stale-price-override correction
green
On 18 September 2026, isolated tests verified that an old credit-setting override stopped governing after the approved price book changed, with the reason made visible. An explicitly reconfirmed setting could then apply under the new version. The record reports production deployment, but states that no override was active, so the deployment itself did not exercise removal of a stale override. The functional proof came from an isolated test identity provider without live fleet or model calls. This record does not define today’s signup-credit amount.
Date: 18 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: No active override during production deploy Functional proof isolated No current credit-price promise
Read the dated evidence summary →Historical missing-signup-credit correction
green
The 18 September 2026 record discloses that some Google signups did not receive the credit advertised on the website. Isolated tests reproduced the failure and then verified that eligible signups received the approved credit even when their identity provider supplied no account-age information. The production section reports deployment, correction of affected Google accounts and clearing of stale low-credit alerts. The automated tests did not use live identity providers or the fleet. This record preserves the customer impact and reported remediation without publishing account identities or balances.
Date: 18 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Some Google signups did not receive the credit advertised on the website. Automated proof isolated Production remediation separately recorded
Read the dated evidence summary →Customer-facing hosted service checks
green
On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.
Read the dated evidence summary →Customer-facing hosted service checks
green
On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.
Read the dated evidence summary →Customer-facing hosted service checks
green
On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.
Read the dated evidence summary →Historical production checkpoint-and-resume canary
green
On 19 September 2026, a synthetic tenant used the production customer routes to create, checkpoint, stop and resume one session. Its memory-backed witness persisted and its counter continued; the completed run cleaned up its test resources. This used one image, one guest size and one resume, with synthetic paid status. It tested the checkpoint route’s orderly stop, not a hard machine kill. Two earlier aborted attempts leaked sessions and were recorded as an open account-deletion issue; the successful cleanup does not erase that limitation.
Date: 19 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: One session/image/size/resume Synthetic paid status Orderly stop, not hard kill Two aborted attempts leaked sessions
Read the dated evidence summary →Historical public-client installation checks
green
Checks recorded on 1 September 2026 verified installation of the released client, successful sign-in and session start on a fresh Linux virtual machine, and rejection of altered installer/update downloads. macOS checks used a cleaned environment rather than a fresh operating-system installation. A 2 September rerun after the repository transfer verified the new distribution paths and compatibility redirects. Windows installation was not covered. These results apply to those historical client releases and environments, not all current installers or supported platforms.
Date: 1 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: macOS environment not fresh VM Windows not tested Historical release paths
Read the dated evidence summary →Historical account persistence and callback tests
green
On 30 August 2026, laboratory checks verified that an account and its session survived a service restart, deletion removed identifying account data while retaining a tombstone, and the deleted account’s session was refused. Altered sign-in callback state created no account. Session-cookie protections were also checked. The login round trip used a test identity provider; Google and GitHub redirect construction was checked without contacting those providers. This is not a live-provider authentication test or proof of document-specific authorization.
Date: 30 August 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Test identity provider Google/GitHub wiring structural only No scoped-document authorization test
Read the dated evidence summary →Historical device-login and token-revocation tests
green
On 30 August 2026, laboratory tests completed the client’s device-login flow, bound the resulting token to the approving account and verified that revocation took effect on the next request. Expired and mismatched approval codes were rejected. The inspected logs and database did not contain the fixture token; it appeared only in the tested client credential store. This exercised the Linux storage path. The macOS keychain path was not tested here, and the result is not a general claim that no token can ever leak.
Date: 30 August 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Linux credential store tested macOS keychain untested here Fixture inspection does not prove universal absence of secrets
Read the dated evidence summary →Historical onboarding and key-vault tests
green
On 30 August 2026, laboratory checks verified that onboarding state survived a restart, stored a fixture provider key in encrypted form and returned only a masked key reference. Cross-account attempts to read or change the vault were refused. The service also refused an unsafe key-storage configuration. The fixture key was deliberately not a real provider credential, and no live provider call was made. The evidence applies to those inspected responses and stores; it does not certify every current production secret-handling path.
Date: 30 August 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Non-provider fixture key No live provider call Inspected paths only
Read the dated evidence summary →Historical usage-meter reconciliation test
green
On 31 August 2026, a test on hosted hardware reconciled a usage statement against its token, session-time and storage records. The total matched its priced lines, an altered meter record failed reconciliation, and an unreadable source was shown as unavailable rather than zero. A real billed model call supplied the token measurement. This test did not exercise card handling or checkout. It verifies the recorded statement and reconciliation checks, not today’s prices or every future bill.
Date: 31 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: No checkout/card path Missing source is unavailable, not zero Historical statement, not current prices
Read the dated evidence summary →Historical credit-ledger and funds-interlock test
green
On 31 August 2026, a hosted-hardware test with a real virtual-machine session verified that repeated credit operations did not duplicate a debit and that exhaustion checkpointed and parked the session, which later resumed after a top-up. A 17 September amendment changed the test because customers were no longer allowed to invoke credit-changing operations. The amended ledger checks run in isolation, with settlement supported by separate evidence. The original assertions describe the earlier interface; they are not proof that those customer operations remain available or that the amended test reran in production.
Date: 31 August 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Original interface superseded Amended ledger proof isolated No production rerun inferred from amendment
Read the dated evidence summary →Historical payment-notification verification test
green
On 31 August 2026, a laboratory sandbox test verified that an authenticated payment notification added credit once, repeated delivery or another notification for the same order added nothing further, and a forged notification was refused. The simulated payment, order and credit records reconciled. A mock payment service and test signing secret were used; this was not a live card purchase or a test of the hosted checkout page. KeepState’s usage statement is separate from the payment provider’s tax receipt. Later live-payment assurance requires its own evidence.
Date: 31 August 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Mock provider and test secret No live card purchase No hosted checkout-page test Usage statement not tax receipt
Read the dated evidence summary →Historical usage-statement integrity test
green
On 31 August 2026, a test on hosted hardware verified that a downloadable usage statement matched its source statement and priced lines, retained a verifiable relationship to the prior statement and detected alteration. It was explicitly a usage statement, not a tax receipt. A 17 September amendment changed how the test account was funded after customer credit-changing operations were closed. This version says the amended rerun was still pending an authorized test window. The original passing result does not establish a completed amended production rerun.
Date: 31 August 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Usage statement, not tax receipt Amended rerun pending in this version Original result not amended production proof
Read the dated evidence summary →Historical low-credit notification test
green
On 31 August 2026, laboratory tests verified that low-credit alerts were calculated from server-side account records, appeared once in the console and were not falsely marked emailed when email was unavailable. With a mock delivery endpoint, an alert reached that endpoint before being marked sent. Live authenticated email was outside this test. A 17 September amendment changed test funding; its attempted rerun was not held because the laboratory environment was unreachable. The original result remains historical, and the aborted attempt is not a passing rerun.
Date: 31 August 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Mock email endpoint Live authenticated email separate Amended rerun trial not held
Read the dated evidence summary →Audited activation and welcome-screen behavior
green
On 31 August 2026, a laboratory test recorded three passing assertions: accounts could not activate themselves, authorized activation was audited, and queued accounts received no start instructions while active accounts received usable welcome instructions. A negative control confirmed that the activation restriction mattered. The welcome test checked instruction structure, not the production destination. This historical result does not establish that the same activation policy or interface is deployed today.
Date: 31 August 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Laboratory run; production URL not verified. Three per-assertion passes, not a product-wide verdict.
Read the dated evidence summary →Approve the task and its checks before running
green
On 16 September 2026, four laboratory assertions passed for Cruise preparation. The client drafted a task manifest, required explicit approval, and refused execution if the approved task or tests changed. A repository without discoverable tests was rejected. Preparation and approval made no service requests in the test. Discovery covered pytest and Go tests only; other test systems were not established by this record. This is a dated client-contract result, not current production verification.
Date: 16 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Only pytest and Go test discovery covered. Laboratory client and recording service; no live execution proved.
Read the dated evidence summary →Retry from the saved state within the approved spending limit
green
On 16 September 2026, a laboratory fixture completed a three-step job after two failed attempts. Each retry started from the same saved state. The test refused an unsupported model choice, stopped when the remaining budget could not cover another attempt, and made no retry after the final attempt. A negative control detected a disabled budget check. These are fixture results, not measured production costs or a savings guarantee.
Date: 16 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Laboratory fixture only No production spending or savings guarantee
Read the dated evidence summary →Each attempt starts from the same saved workspace
green
On 17 September 2026, six infrastructure assertions passed for isolated attempts from one saved workspace, restored content, periodic save points and cleanup on hosted infrastructure. The addendum records that every attempt failed: the test accounts lacked the required provider key, so no inference call or task acceptance occurred. Successful-attempt parking was conditional and was not demonstrated. This one-job run supports the tested infrastructure behavior, not task success or acceptance rates. Checkpoint pauses remain host-specific measurements.
Date: 17 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: No inference call or task acceptance: test accounts had no provider key. Successful-attempt parking was conditional and not demonstrated. One job; checkpoint pauses are host-specific.
Read the dated evidence summary →Independent verification rejects a forged pass
green
On 16 September 2026, a hosted-service test recorded four passing checks: the candidate was extracted and the child stopped before verification; the verifier ran on the host using approved tests; verdicts carried signed test references; and a forged pass log was rejected. A negative control showed that compromised verifier independence could invalidate the result. The tested job ended in review after its single attempt failed. This establishes the tested verification boundary, not universal task success or current deployment state.
Date: 16 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: One hosted attempt The job ended in review, not acceptance Current deployment independently unverified
Read the dated evidence summary →Unresolved attempts remain visible in the job record
green
On 16 September 2026, four laboratory assertions passed for job accounting. Attempts progressed through their recorded lifecycle, reported spend matched reconciled attempts, and the event history passed its integrity check. When usage remained unresolved, the attempt stayed visible and the job could not be accepted. A negative control caught an attempted acceptance despite that ambiguity. These checks used a fixture service with a simulated reconciliation timeout; they did not measure a live provider’s reporting delay.
Date: 16 September 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Fixture backend and modeled timeout. No live-provider reconciliation-latency claim.
Read the dated evidence summary →Job status and recorded results
green
The original laboratory record is dated 16 September 2026. A 17 September amendment changed test funding after customer credit grants were disallowed; the revised normal and negative-control runs each passed eight checks. They rejected a changed approval digest and a second concurrent free-tier job, verified event history, prevented acceptance of unresolved attempts and resume outside review, and displayed missing spend as unavailable. Customers could not grant themselves credit. These local tests used a test identity service and do not establish current production limits or availability.
Date: 16 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Laboratory services and test identity provider. Customer credit changes require authorized support. Historical tier constraint requires current product confirmation.
Read the dated evidence summary →Offline help and verified artifact downloads
green
On 16 September 2026, five laboratory assertions passed for the Cruise client. Twenty help invocations made no service requests, status and logs preserved the recorded states and event order, and a modified artifact was refused before writing an output file. Workspaces larger than 50 MiB were rejected before upload; the measured size excluded Git and KeepState metadata directories. This record covers the tested client contract and recording service, not every current client release.
Date: 16 September 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: 50 MiB threshold and exclusions belong to this tested release. Recording service, not full production integration.
Read the dated evidence summary →A hosted job through the customer-facing service
green
On 17 September 2026, seven assertions passed in a production-hosted check with a bring-your-own-key account. A job was accepted, an unrelated test tenant could not see it, its event history verified, and the downloaded artifact matched its recorded digest. A successor run that day repeated the checks after client corrections. Monetary spend remained unavailable: usage records did not include money, so zero was not a valid cost claim. Acceptance was an observed outcome, not a required success rate. Artifact verification was conditional on acceptance, and the record does not establish other workloads or later deployments.
Date: 17 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Monetary spend unavailable; usage does not establish money spent. Limited job/account scenario; acceptance is not a performance promise. Artifact verification applies only to accepted runs.
Read the dated evidence summary →Review, resume and cancellation after failed attempts
green
On 16 September 2026, four fixture-based laboratory assertions passed. Exhausting the approved attempt plan moved a job into review with a notification event. Extending the plan from review allowed the tested job to complete; cancellation released the retained workspace. The final permitted attempt was not silently retried. Resume was allowed only from review, and the retained parent remained the customer’s recovery option until release. This record does not prove live notification delivery or production availability.
Date: 16 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Fixture backend. Resume only from review. Notification event is not proof of message delivery.
Read the dated evidence summary →Email is recorded before delivery is claimed
green
The original record dated 2 September 2026 reports three passing assertions across laboratory and hosted email checks. Simulated interruptions left mail queued, and sent status required provider acceptance. The provider verified the domain’s email authentication configuration and accepted a welcome message. The interruption tests used a mock provider; provider acceptance does not prove inbox delivery or receiver authentication headers. A 17 September amendment changed test funding, but its rerun was not held because the laboratory host was unreachable. The amendment therefore adds no new passing run.
Date: 2 September 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Mock provider for interruption tests. Provider acceptance does not prove recipient receipt or receiver headers. 17 September amended rerun was not held because the laboratory host was unreachable.
Read the dated evidence summary →Production transactional email sending check
green
On 9 September 2026, a production transactional-email check recorded a green result. The provider reported the sending domain authenticated, a DMARC policy was present in public DNS, and the deployed service’s message was accepted by the mail provider. The check did not independently derive every DNS authentication record or measure inbox placement. It did not establish delivery to a subscriber list; subscriber fan-out and an unsubscribe route were absent in this record.
Date: 9 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Provider authentication-state reporting, not independent DNS derivation. No inbox-placement or subscriber-delivery proof. Subscriber fan-out and unsubscribe route absent.
Read the dated evidence summary →Account erasure and recovery isolation
green
On 7 September 2026, six recorded assertions passed for account-level erasure on hosted infrastructure. Deleted-account checkpoints became unreadable after key destruction, including from the tested backup copy, while another account’s matching content remained restorable. Identity was removed from retained billing records. The rollout used team/test data; a cleanup error also removed the test canary’s chunks, with no reported customer impact. A fresh post-cleanup checkpoint subsequently resumed. This record does not establish that every backup or future deployment follows the same key-management conditions.
Date: 7 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Team/test rollout only; cleanup error and no reported customer impact retained. Backup recovery depends on stated key-management conditions.
Read the dated evidence summary →Deletion fails closed when erasure cannot be confirmed
green
On 7 September 2026, four assertions passed using production binaries with disposable data. The customer deletion path removed credentials and made checkpoint data unrecoverable through the tested backup-recovery path. It did not report deletion if key destruction could not be confirmed. A second tenant remained restorable. Production services were not changed by this test; it exercised their deletion code in isolation. Recovery protection depends on keys being excluded from the restored backup.
Date: 7 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Disposable data; production daemons untouched. Backup/key separation is required.
Read the dated evidence summary →Deletion verified on the running hosted service
green
On 8 September 2026, three production-hosted assertions passed for deletion of a synthetic tenant. The customer-facing deletion operation destroyed checkpoint access, removed stored provider credentials and removed the account identity without affecting another tenant. A copied checkpoint was unrecoverable under the stated key-management conditions. The fix was deployed during this work; the result applies from that deployment, not earlier builds. The copy was made for the test because no scheduled checkpoint-store backup existed.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenant. Attests only from the corrected deployment onward. Test-made backup copy; no scheduled checkpoint-store backup.
Read the dated evidence summary →Deleted credentials stay unreadable in the tested backup
green
The credential-erasure record dated 8 September 2026 reports three passing hosted test assertions. A database backup recovered the test credential before account deletion but could not afterward, even with the master key retained. A negative control confirmed that destroying the account key was necessary. The base test required older entries to be rewritten; an appended production section dated 7 September reports migrating and verifying all four existing entries. Erasure protection depends on keeping account keys outside the database backup set. The record does not establish later deployments or independent copies of those keys.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Separate account-key custody and exclusion from the database backup set are required. Base test and production migration appendix are distinct evidence. Later deployments and independent key copies are outside this record.
Read the dated evidence summary →Credential erasure protection enabled in production
green
The production record dated 8 September 2026 reports enablement on 7 September. Its single assertion passed: all four existing credential entries were migrated to per-account protection and verified to require their account keys rather than the master key alone; a live model request still succeeded. This supports the reported migration at that time. Credential erasure depends on keeping account keys outside the database backup set. The record does not certify later entries, later deployments or independent copies of key material.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Erasure depends on account keys being excluded from database backups. Historical migration result, not continuing certification. 8 September record date differs from 7 September reported enablement.
Read the dated evidence summary →Resume on another host after an interruption
green
On 5 September 2026, four hosted assertions passed for moving saved sessions to another host after an isolated service failure. The tested sessions preserved in-memory state, each move produced one event and a customer notice, and duplicate resume attempts were prevented. The test used shared durable storage within one region. Work since the latest checkpoint could be lost. Billing reconciliation on the replacement host was still pending in this record; these results do not prove recovery from every host or regional failure.
Date: 5 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Isolated service termination, same-region shared storage. Checkpoint-bounded loss window. Replacement-host billing reconciliation pending in this record.
Read the dated evidence summary →Billing state across a real host loss
green
The record combines a hosted failover rehearsal dated 6 September 2026 and production enablement dated 7 September. Six assertions report billing continuity after real host deallocation: recorded spend survived, an identical retry was not billed again, and budget enforcement used the carried balance. Settlement stayed blocked until reconciliation passed. One run measured 47 seconds from deallocation to reconciliation; this is not an availability promise. The enablement addendum records a live production check. Session recovery relied on a separately tested mechanism, and regional disaster recovery was not established.
Date: 6 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: 47 seconds is one run’s deallocation-to-reconciliation measurement. Shared storage and separately tested guest recovery are dependencies. Settlement remains blocked until reconciliation completes. Production enablement is a separate 7 September addendum.
Read the dated evidence summary →Customer-facing hosted service checks
green
On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.
Read the dated evidence summary →Historical environment-readiness test
green
On 28 July 2026, the laboratory environment passed readiness checks for running hardware-virtualized guests and verifying its pinned software. A deliberately altered integrity value was rejected. The result supports that environment’s readiness and integrity checks at that point in time. It does not demonstrate an agent workload, hosted availability or ongoing production health.
Date: 28 July 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Point-in-time laboratory hygiene Single altered integrity value No hosted availability claim
Read the dated evidence summary →Historical memory-continuity test
green
On 28 July 2026, a laboratory guest resumed after its virtual-machine process was forcibly stopped. Its in-memory witness remained the same and its counter continued advancing; a fresh boot failed the continuity checks. The guest used a read-only filesystem, so the witness could not be reconstructed from a saved disk file. This was a small counter guest, not an AI agent or a production workload, and establishes only the recorded memory-continuity result.
Date: 28 July 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Counter guest, not AI agent Read-only filesystem test Laboratory, not production
Read the dated evidence summary →Historical session lifecycle test
green
On 28 July 2026, laboratory checks ran a session through checkpoint, forced stop and resume while preserving its in-memory witness and continuing counter. Repeating the test confirmed the lifecycle could run again, and substituting a fresh start failed the continuity checks. This test did not verify clock resynchronization; a separate general-image demonstration was reported as data, not an asserted result. The evidence concerns the recorded laboratory lifecycle, not current hosted availability.
Date: 28 July 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Clock resynchronization untested General-image smoke result is data only Laboratory scope
Read the dated evidence summary →Historical checkpoint portability and integrity test
green
On 28 July 2026, a laboratory test restored a session from its stored checkpoint on a second service instance, created branches with recorded ancestry, and rejected corrupted checkpoint data. An interrupted checkpoint publication left the prior checkpoint intact. Both service instances ran on the same laboratory host with compatible software and local storage; this was not a cross-host disaster-recovery test. Branch differences came from staggered execution, not agent steering.
Date: 28 July 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Second instance on same host Compatible software/local storage Not cross-host recovery Branch divergence not steering
Read the dated evidence summary →Historical interrupted-call replay test
green
On 28 July 2026, a laboratory Claude Code session survived a forced stop and resume. Its records reconciled, and a completed interrupted response was replayed without being billed again. The test also detected deliberately degraded replay and record behavior. The result depends on the client returning for the held response; it is not a promise about every agent’s retry behavior. The specimen’s task outcome was reported separately and did not determine whether the infrastructure test passed.
Date: 28 July 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Claude Code client scope Client retry behavior separate Task outcome not asserted Some billing checks alone remain true in degraded system
Read the dated evidence summary →Historical unattended recovery demonstration
green
On 29 July 2026, a laboratory demonstration automatically checkpointed an agent, stopped both its virtual machine and supervising service, and resumed the session. The recorded run passed its recovery and cleanup checks. A run without a usable checkpoint correctly failed. The asserted endpoint was the agent being alive after resume; completion of its task was separate reported data. Reported component timings were measurements, while monetary conversions were estimates. This is a single demonstration result, not a general recovery-time or task-success guarantee.
Date: 29 July 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Alive after resume, not task completion Component timings reported, not all asserted Dollar figures estimated Single laboratory demonstration
Read the dated evidence summary →Historical branch-isolation and steering test
green
On 29 July 2026, laboratory checks created branches from a shared checkpoint, applied distinct steering inputs and verified separate branch histories and filesystems. A test-written marker in one branch did not appear in the others. Identical steering inputs were correctly detected as failing the distinctness requirement. The marker was written by the test harness, not the agent: it proves filesystem isolation, not that an agent obeyed its instructions. Agent behavior remains measured data rather than a guaranteed result.
Date: 29 July 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Marker written by test harness Isolation is not agent obedience Laboratory scope
Read the dated evidence summary →Historical long-running checkpoint reliability test
not-required
The snapshot records this historical soak as not required for the release rerun; its old result is not a fresh run. The record sealed on 31 July 2026 reports a replacement seven-hour laboratory soak with successful scheduled checkpoints and restoration of an older checkpoint. An earlier run exposed a defect that stopped repeated checkpointing; the record reports its correction and retest. It also discloses that cleanup destroyed the originally preserved checkpoint, so its restoration was not performed; the later restoration used a replacement artifact. One workload and one soak do not establish a general reliability rate. The separate 25-cycle sample supports only its stated confidence interval, not guaranteed success.
Date: 31 July 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: One soak/one workload Original preserved artifact destroyed; original restoration not performed Replacement restoration is distinct 25-cycle confidence limits remain visible
Read the dated evidence summary →Historical second-agent compatibility test
green
The original record of 31 July 2026 and rerun of 30 August tested a second agent, aider, and a fresh laboratory installation. The gateway retained the interrupted response in all four reported runs; aider retried successfully in three. That difference is a material client limitation. The second installation used another virtual machine on the same physical computer. A later attempt was not counted as a completed trial because its environment was not ready. These results support the named agent and setup, not compatibility with every agent or guaranteed resumption.
Date: 31 July 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Aider retry 3/4 versus held response 4/4 Second VM on same physical machine Trial-not-held excluded explicitly Not universal compatibility
Read the dated evidence summary →Customer-facing hosted service checks
green
On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.
Read the dated evidence summary →Customer-facing hosted service checks
green
On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.
Read the dated evidence summary →Separate recovery, metering and retry support checks
green
On 5 September 2026, eight hosted assertions passed across two supported provider paths. Test sessions preserved in-memory state through save, interruption and resume; real provider calls were metered; identical retries avoided a second charge; missing credential access failed closed. The tests used a base guest exercising the integration configuration, rather than proving every behavior of a full agent application. Whether a request is retried remains the client’s decision. These results apply to the tested combinations.
Date: 5 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Base guest/configuration surface, not every native agent workflow. Retry issuance is client-controlled.
Read the dated evidence summary →Separate recovery, metering and retry support checks
green
On 5 September 2026, eight hosted assertions passed across two supported provider paths. Test sessions preserved in-memory state through save, interruption and resume; real provider calls were metered; identical retries avoided a second charge; missing credential access failed closed. The tests used a base guest exercising the integration configuration, rather than proving every behavior of a full agent application. Whether a request is retried remains the client’s decision. These results apply to the tested combinations.
Date: 5 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Base guest/configuration surface, not every native agent workflow. Retry issuance is client-controlled.
Read the dated evidence summary →Protect unsaved work from accidental termination
green
On 31 August 2026, three hosted assertions passed for the termination guard. An ordinary termination request with unsaved work was refused and the session stayed alive, with safer choices explained. Explicit forced termination proceeded; a session with a saved checkpoint could be terminated normally. The protection was enforced by the service as well as the client. This guard protects unsaved work, not all destructive actions, and an explicit force request can still discard that work.
Date: 31 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Guard concerns unsaved work. Explicit force permits destruction.
Read the dated evidence summary →Historical accepted-work test across two provider families
green
A confirmation run on 6 September 2026 completed one bounded coding fixture through each of two live provider families. The initial attempts failed; later attempts produced accepted artifacts after independent checks. Seven checks passed, including complete-job accounting. Publication was limited to a scoped authorization: no external pull request was created. A negative control showed that verifier independence mattered. The sample is two task arms with deliberately constrained initial attempts, not a general task-success rate, comparative model ranking or cost-savings claim.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: One bounded fixture in two provider families Initial attempts were deliberately resource-constrained No external publication or general success rate
Read the dated evidence summary →Rehearsal of onboarding, usage and billing reconciliation
green
On 31 August 2026, a hosted rehearsal passed all ten steps in the recorded customer journey and two summary assertions. It covered activation, a supplied provider key, crediting, model use, pause/resume, statement reconciliation and a refund returning the balance to zero. Revoked access and a modified statement were rejected. The identity provider and payment service were test systems: this was a rehearsal, not proof that a real new customer completed live checkout or received a real refund.
Date: 31 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Test identity provider and mock payment sandbox. No live customer checkout or live refund established.
Read the dated evidence summary →Metering and replay on an additional provider path
green
On 4 September 2026, six hosted assertions passed for an additional compatible provider path. Deterministic tests checked credential routing, usage accounting, ordered records and replay of an identical request without another charge. A separate live provider call also succeeded and its identical retry was replayed. The live gateway ran in isolation and did not prove the full agent-plus-session-interruption combination. The replay guarantee concerns identical retries; it does not mean all similar requests are free.
Date: 4 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Mechanics used a deterministic mock. Live gateway was isolated; full agent recovery separate. Replay applies to identical requests.
Read the dated evidence summary →Change models only while a session is saved
green
On 4 September 2026, three hosted test assertions passed for changing a session’s model. A change was refused while the session was running, accepted while it was checkpointed, and used by the next request without resetting the request lineage. These tests used a simulated compatible provider. They covered compatible clients through the tested routing path; native clients for the other protocol were explicitly outside this result. Live continuation quality was not measured here.
Date: 4 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Mock upstream. Native clients for the other protocol not claimed. No live continuation-quality measurement.
Read the dated evidence summary →Continue a fixture across provider protocols
green
On 7 September 2026, four hosted assertions passed for continuing a tested task across provider protocols at a checkpoint boundary. The live continuation included tool use and an independently verified artifact, while lineage and metering remained continuous. A failed continuation was refused and the prior workspace restored for fallback. Deterministic translation checks were separate from the live run. The result is fixture-specific and model-dependent; streaming translation was unsupported and explicitly refused in this version.
Date: 7 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Fixture-specific continuation; quality depends on model. Streaming unsupported in this version. Deterministic and live checks have separate scopes.
Read the dated evidence summary →Production check of cross-provider translation
green
On 8 September 2026, one production-hosted assertion passed for a real translated request on the team’s account. The request reached a live compatible model, was recorded and metered, and the service stayed healthy. This was a one-call production check, not a broad task-success study. Streaming responses were buffered rather than incremental, and replayed thinking blocks were omitted during translation. Those limitations remain part of the result; current behavior requires separate deployment verification.
Date: 8 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: One call on a team account. Buffered streaming; replayed thinking blocks omitted.
Read the dated evidence summary →Native coding-agent continuation across providers
green
On 7 September 2026, three hosted assertions passed with the actual tested coding-agent application and real tool use. A fixture task continued after a checkpoint-boundary provider change and produced an independently verified artifact. Lineage and metering remained continuous; a disabled continuation path failed closed and fallback completed after restoring the workspace. The agent ran in an isolated copy of its environment, with transport tested separately. Quality remains model-dependent, and streaming was buffered rather than incremental.
Date: 7 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Isolated application environment; transport proved separately. Fixture and model-dependent result. Buffered, nonincremental streaming.
Read the dated evidence summary →Independent quickstart run for release v0.1.3
green
On 25 August 2026, a clean-environment run of the v0.1.3 quickstart completed successfully on an Apple Silicon laboratory machine. The source instructions were hash-checked before execution. The first attempt stopped because a stale checkout existed; a fresh second attempt exited successfully in 176 seconds, preserved session state across resume and completed cleanup. This was a quickstart demonstration, not certification of the separate blocked acceptance test. That broader test remained blocked on two assertions awaiting a human decision.
Date: 25 August 2026
Scope: Laboratory or bench evidence; production behavior is not inferred from it.
Limit: Two attempts; stale-checkout refusal retained. 176 seconds belongs to one successful attempt. Broader acceptance gate remained blocked.
Read the dated evidence summary →Measured session resume times
green
On 5 September 2026, a hosted measurement ran 20 warm and 20 cold resumes of a 2 GiB guest in interleaved order. All 40 completed. Warm p95 was 25.09 seconds and cold p95 was 47.46 seconds; medians were 18.08 and 33.36 seconds. The cold result met the 60-second target, while the warm result missed the 5-second target. These measurements describe that guest and environment, not a general latency guarantee or current service level.
Date: 5 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Warm p95 missed the target. Single tested guest/environment and 40 runs. Measurement, not service guarantee.
Read the dated evidence summary →Deterministic checks for the acceptance contract
green
On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.
Read the dated evidence summary →Accepted-work checks on a plain Linux stand-in
green
On 7 September 2026, three assertions passed for the accepted-work workflow on a plain Linux customer-environment stand-in. A failed attempt restarted from a saved segment boundary, a later attempt completed, and the artifact was independently checked after the remote round trip. The environment’s limitations appeared in the result and missing limitation declarations were rejected. Plain-host recovery did not rewind memory. This was a tested integration configuration, not proof for every customer host or workload.
Date: 7 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Customer-environment stand-in. No RAM rewind on plain hosts; segment restart only. Configuration-specific result.
Read the dated evidence summary →Deterministic checks for the acceptance contract
green
On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.
Read the dated evidence summary →Live task completion after a failed attempt
green
On 6 September 2026, four hosted assertions passed across two live provider families. A first-stage candidate could pass independent checks; when a first attempt failed, the workspace was restored and a later stage completed the fixture. Calls were metered, and disabling escalation did not turn failure into acceptance. The tasks were tiny supplied fixtures, and recovery here was at workspace-segment level. Memory restoration was established separately; no acceptance-rate, cost-saving or general workload claim follows.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Tiny supplied fixtures. Workspace-level segment restore; memory recovery separate. No savings or general acceptance-rate claim.
Read the dated evidence summary →Deterministic checks for the acceptance contract
green
On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.
Read the dated evidence summary →Deterministic checks for the acceptance contract
green
On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.
Read the dated evidence summary →Deterministic checks for the acceptance contract
green
On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.
Read the dated evidence summary →Deterministic checks for the acceptance contract
green
On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.
Read the dated evidence summary →Self-service signup: status observation and team attestation
green
On 14 September 2026, the public production status reported self-service signup enabled and operational. Separately, the QA team attested that real-provider signups had completed, automatically activated accounts and received the signup credit during the preceding week. These are different evidence types: the status observation proves the setting was enabled; the successful-signup claim is a team attestation, not an independently driven browser test in this record. Current signup behavior requires a fresh end-to-end check.
Date: 14 September 2026
Scope: Historical record; any production observation is limited to the dated test described below.
Limit: Enabled status does not prove signup completion. Successful signup is team attestation, not a driven canary.
Read the dated evidence summary →Reissued six-hour checkpoint evidence
green
The reissued record spans the 31 August 2026 release and 7 September test evidence. It preserves the original hosted segment’s failure to reach 72 checkpoints: only 68 were recorded, although recovery of old state and bounded memory passed. A later run using the corrected driver reached 72 checkpoints across its own six hours and restored the earlier checkpoint. Reissue did not turn the first result into a pass. This is evidence for two identified test segments, not a guarantee for every workload or current release.
Date: 31 August 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Original failed segment retained. Later corrected run is separate. Record spans release and run dates.
Read the dated evidence summary →Encrypted checkpoint soak: one checkpoint missed
red
The encrypted soak failed because one expected checkpoint was missed. The later rerun is a separate record; this failed result stays visible. The run completed in September 2026 and was sealed as failed on 8 September. On the encrypted hosted configuration, a six-hour segment restored its hours-old state and stayed within the other measured durability limits, but only 71 of 72 checkpoints completed because one snapshot exceeded its deadline. The record therefore remains failed. A later amendment points to a successful rerun completed on 19 September; that result belongs to the separate rerun and does not erase this defect or its original outcome.
Date: 2026-09-08
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: 71/72 failure preserved. Run transcript dated 7 September; failed seal dated 8 September. Later supersession is not a revised original pass.
Read the dated evidence summary →Six-hour encrypted checkpoint rerun completed
green
From 18 to 19 September 2026, one hosted six-hour segment completed all 72 checkpoints after the earlier deadline defect was corrected. Six recorded assertions passed, covering the segment duration, checkpoint count, cadence, bounded memory and absence of checkpoint failures, plus a control against pooling unrelated runs. This rerun used one guest size on one host. It does not turn the earlier failed run into a pass or establish reliability for larger guests, other workloads or the current deployment.
Date: 19 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: One guest size and one host. Earlier failed run remains failed. Date taken from body because viewer metadata is incomplete.
Read the dated evidence summary →Evidence catalogue titles and source records
green
On 6 September 2026, two deterministic publication assertions passed. Evidence titles and summaries were derived from each record’s declared subject, conflicting manually supplied titles were rejected, and records missing required metadata stopped generation. This checks the consistency of the evidence catalogue, not the truth of every underlying product claim. It covers the record-generation side; whether the public website actually rendered the correct titles required a separate website check.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Catalogue consistency, not product verification. Public-site rendering tested separately.
Read the dated evidence summary →Public claims and their recorded evidence
green
On 6 September 2026, three deterministic publication assertions passed. Claimed results had to reference specific assertions present in the cited records, rather than just naming a test or release. The check reconciled 152 assertions across 42 test records and rejected empty or invalid references. This establishes traceability within the inspected records, not whether each product result is correct or currently deployed. The public website’s rendering was a separate scope.
Date: 6 September 2026
Scope: Historical evidence; the scope and environment are described in the recorded result.
Limit: Record linkage, not independent product correctness. Website rendering separate.
Read the dated evidence summary →Agent-session production canary on 30 September 2026
green
On 30 September 2026, a production canary created an agent session, bound the account’s own provider key, started its supervised machine and completed an instruction with Claude Code 2.1.251. Eight normal-run checks passed. A separate negative check refused a model call without a bound key. Usage showed a metered call; dollar cost remained unavailable because no approved model rate existed. Deletion stopped the tested machine, while checkpoint content remained subject to retention. An earlier run printed an empty runner label; a corrected final run followed. This one-account demonstration does not validate every agent, adviser, budget, approval or recovery workflow.
Date: 30 September 2026; final normal run at 09:16:29 UTC
Scope: One team-operated production account using its own provider key and Claude Code 2.1.251.
Limit: One production account, one supported runner version and a bounded instruction. Eight normal checks and one no-key negative check; model reply content was not the success criterion. The registry declared availability for several features; the canary did not independently exercise every declared workflow. Dollar cost was unavailable because no approved rate was recorded. Session deletion stopped the tested machine; checkpoint-content erasure was not tested. The first run’s runner-label display was corrected before the final normal run; its original pass conditions did not depend on that label. Open findings about approval timing, report integrity and unpriced spending are not closed by this canary.
Read the dated evidence summary →The public summary as JSON. These are the same reviewed fields shown in Read.
{
"schemaVersion": 1,
"publicSummaryVersion": "2026-09-30",
"classification": "PUBLIC",
"title": "Test-status register updated 30 September 2026",
"summary": "The 30 September 2026 register contains 98 test records: 96 marked green, one failed and one not requiring a rerun. It adds the dated agent-session production canary. The failed encrypted soak remains recorded even though its stated later rerun succeeded; a not-required entry remains distinct from a passing rerun. Each row retains its historical test date and scope. This updated register is not a new execution of all 98 tests, a universal agent-compatibility result, or independent verification of today’s deployment.",
"date": "30 September 2026 snapshot; individual historical test dates are shown.",
"scope": "Published register or policy snapshot; per-item scope and dates apply.",
"result": "The 30 September 2026 register contains 98 test records: 96 marked green, one failed and one not requiring a rerun. It adds the dated agent-session production canary. The failed encrypted soak remains recorded even though its stated later rerun succeeded; a not-required entry remains distinct from a passing rerun. Each row retains its historical test date and scope. This updated register is not a new execution of all 98 tests, a universal agent-compatibility result, or independent verification of today’s deployment.",
"sections": [
{
"title": "Individual recorded test outcomes",
"items": [
{
"id": "gate-1b-1",
"title": "Historical opt-in encrypted-record test",
"text": "On 7 September 2026, an engine test on hosted hardware verified that an opted-in session produced an encrypted transcript readable with the customer’s key, while an opted-out session produced no transcript in the record subsystem. Alteration of a recorded entry was detected. This record explicitly says the feature was not enabled in production. The separate, bounded response-replay cache is outside the transcript subsystem, so the opt-out result is not a claim that no response bytes can exist anywhere. Production enablement requires separately reviewed privacy disclosures and evidence.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Not enabled in production in this record Replay cache outside recording subsystem Engine test, not current deployment proof",
"href": "/evidence/phase-1b-1"
},
{
"id": "gate-1b2d",
"title": "Historical transcript-conversion test",
"text": "On 7 September 2026, deterministic checks converted a real Claude Code transcript into inputs for other agent tools and produced a separate task handoff. The record reports the conversation information lost in each conversion. No live target agent was run in this test, and no successful task continuation is claimed. Each supported source/target pair needs its own live completion test before it can be advertised as harness switching. The task handoff remained usable independently of transcript conversion.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Mechanics only; no live target agent Only one source-transcript reader tested Conversation loss declared per target No harness-switch availability claim",
"href": "/evidence/phase-1b2d"
},
{
"id": "gate-ac-1",
"title": "Historical signup-credit and free-session test",
"text": "On 1 September 2026, laboratory tests with real virtual-machine sessions verified automatic account activation, a one-time signup credit, concurrency limits and checkpoint-and-park behavior. Repeat signup by the same authenticated identity did not receive another grant. The record’s parking behavior was based on elapsed runtime, not inactivity, and some abuse controls were approximations. Production self-service was disabled in this historical version. A 17 September amendment changed the credit-exhaustion check to an isolated internal test; that amendment is not itself a new full production run. This record does not establish current signup grants or availability.",
"status": "green",
"date": "1 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Runtime rather than idle detection Abuse-control approximations Historical production self-service disabled Amended exhaustion test isolated",
"href": "/evidence/phase-ac-1"
},
{
"id": "gate-ad-1",
"title": "Historical staff-console control tests",
"text": "On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.",
"status": "green",
"date": "1 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.",
"href": "/evidence/phase-ad"
},
{
"id": "gate-ad-2",
"title": "Historical staff-console control tests",
"text": "On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.",
"status": "green",
"date": "1 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.",
"href": "/evidence/phase-ad"
},
{
"id": "gate-ad-3",
"title": "Historical staff-console control tests",
"text": "On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.",
"status": "green",
"date": "1 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.",
"href": "/evidence/phase-ad"
},
{
"id": "gate-ad-4",
"title": "Historical staff-console control tests",
"text": "On 1 September 2026, four recorded tests on hosted hardware reported role restrictions, auditing of privileged writes, consent-bound content access, independently reconciled financial displays and detection of altered records. Missing source data was displayed as unavailable rather than zero. These are historical test assertions about the administrative console. They do not establish the current confidential-document access model or resolve conflicting public statements about whether staff-access controls are implemented. This historical record does not independently verify the deployed staff-access behavior.",
"status": "green",
"date": "1 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Historical console tests No document-pack access proof The historical record contains conflicting staff-access statements.",
"href": "/evidence/phase-ad"
},
{
"id": "gate-af-1",
"title": "Historical hosted attach-and-fork test",
"text": "On 3 September 2026, laboratory checks created owned fork branches, exercised terminal attachment and refused another account’s access. File-divergence checking was skipped because most branch networks were unreachable, and attachment used the stable parent. A separate production narrative reports successful owned forks and keyboard interaction through the hosted client after deployment. Final robustness changes could not be rerun on the damaged laboratory environment. The laboratory assertions and production narrative cover different scopes; neither resolves the recorded branch-network limitation.",
"status": "green",
"date": "3 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "File-divergence check skipped in laboratory Attachment tested on stable parent Separate production narrative Final robustness changes not rerun on damaged laboratory environment",
"href": "/evidence/phase-af-1"
},
{
"id": "gate-af-2",
"title": "Historical independently verified branch-work test",
"text": "On 7 September 2026, a hosted-hardware test forked one task workspace into three branches. Two branches reached authorized model providers, performed work in their own workspaces and produced accepted candidates under independent verification. The unreachable branch was counted as failed rather than omitted. Exactly one verified result received scoped publication authorization. This was authorization for a review branch or pull request, not an actual external publication. Larger branch counts were not tested.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Three branches only Unreachable attempt counted failed Authorization intent, no external publication",
"href": "/evidence/phase-af-2"
},
{
"id": "gate-az-0",
"title": "Historical deployment-platform portability test",
"text": "The record resolved on 30 August 2026 reports that the existing readiness, memory-continuity, lifecycle and checkpoint tests passed on the hosted x86 platform. Their test definitions were checked for consistency with the laboratory versions. The first run failed because its corruption test did not actually change the selected data; after that test was corrected, the negative test passed. This demonstrates the recorded platform transition, not continuous availability, portability across arbitrary architectures or a claim that the first failed result never occurred.",
"status": "green",
"date": "30 August 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Initial corruption test invalid, corrected in later run Named platform only No arbitrary cross-architecture portability",
"href": "/evidence/phase-az-0"
},
{
"id": "gate-az-1",
"title": "Historical account-isolation test",
"text": "On 30 August 2026, a two-account test on hosted hardware verified that each owner could operate its own session and that forged or cross-account attempts to read, attach, checkpoint or resume were refused without changing the target session. The test exercised the engine’s account-ownership boundary using test-generated account assertions. It was not a full end-to-end test of external sign-in or today’s document portal. Legacy laboratory sessions had a separate local-access scope, explicitly outside the owned-session claim.",
"status": "green",
"date": "30 August 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Two-account engine test Not external sign-in or document authorization Legacy laboratory-session scope differs",
"href": "/evidence/phase-az-1"
},
{
"id": "gate-az-2",
"title": "Historical owned-session broker test",
"text": "On 30 August 2026, a test on hosted hardware used an account token to create, read and checkpoint an owned session through the control plane. Another account’s token and a revoked token were refused. The tested components were colocated; public-domain transport, real Google/GitHub sign-in and real provider-key use were not established by this test and awaited separate setup. The result demonstrates that recorded ownership and token-revocation path, not a complete customer onboarding or payment journey.",
"status": "green",
"date": "30 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Colocated components Real external sign-in and BYOK outside scope Not complete paid onboarding",
"href": "/evidence/phase-az-2"
},
{
"id": "gate-az-3",
"title": "Historical private service-boundary test",
"text": "The record sealed on 31 August 2026 reports checks run on 30 August showing that the authorized control plane could create and read an owned session across the private network. Public and unauthenticated transport attempts were refused, including plaintext and untrusted client certificates. Network restrictions and certificate authentication were tested independently. This was a same-region control-path test. It did not test moving a checkpoint to another host, recovery after host loss or today’s network configuration.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Run date differs from seal date Same region control path Not cross-host restore Historical configuration only",
"href": "/evidence/phase-az-3"
},
{
"id": "gate-az-3b",
"title": "Provider-key isolation and safe refusal without a key",
"text": "On 30 August 2026, a hosted-service test used a fixture credential and a simulated provider. The session completed its request using the account’s supplied key, the inspected logs and stored records contained no plaintext copy, and a session without a usable key was refused. A negative control confirmed refusal when the key could not be decrypted. Assertions were appended on 6 September 2026. This record did not make a live provider call and does not establish that arbitrary credentials can never leak.",
"status": "green",
"date": "30 August 2026; 6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Simulated provider and fixture credential Live provider behavior belongs to a separate record No general proof against credential leakage",
"href": "/evidence/phase-az-3b"
},
{
"id": "gate-az-4",
"title": "Historical hosted-client and provider-key trial",
"text": "The record sealed on 31 August 2026 reports hosted-client checks run on 30 August: a token created a session on the live fleet, read its usage information and stopped it. A revoked token was refused before creating a session. A separate team-operated trial on 31 August completed a real model call using the team’s own provider key. That real-key trial is narrative evidence, not one of the automated gate assertions. These two scopes are distinct and do not establish behavior for every workload or current release.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Automated client checks versus narrative real-key trial Team-owned key, not customer trial Historical availability only",
"href": "/evidence/phase-az-4"
},
{
"id": "gate-az-5",
"title": "Historical restart-and-reboot recovery test",
"text": "On 31 August 2026, a hosted-hardware test verified that the services restarted after a forced process stop and returned automatically after the host rebooted. A previously checkpointed and stopped counter session resumed with its in-memory witness preserved and its counter advancing. The negative test detected the absence of automatic restart. This test used a small counter specimen and a saved checkpoint; it does not promise preservation of unsaved work, arbitrary workload recovery or failover to another host.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Counter specimen Checkpointed and stopped before reboot Not cross-host failover or unsaved-work guarantee",
"href": "/evidence/phase-az-5"
},
{
"id": "gate-b37",
"title": "Historical erasure-failure isolation regression",
"text": "A regression run on 7 September 2026, sealed in the 8 September record, verified that a session with destroyed account key material could not resume while another account’s session resumed immediately afterward without restarting the service. The earlier reported service failure was traced to an invalid test request, not damage to the service. This was an isolated engine test on hosted hardware. It did not exercise the production account-deletion route, which requires its separate canary evidence.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Earlier failure was invalid test request Isolated engine test Production deletion route separate",
"href": "/evidence/phase-b37"
},
{
"id": "gate-b38",
"title": "Historical account-deletion cleanup correction",
"text": "On 8 September 2026, regression and production-canary evidence reported that deleting a synthetic account removed its active-session resources before destroying its key. The prior behavior left unusable session residue and caused health checks to fail. The corrected deletion refused to report success if teardown failed. Billing-related tombstone records were retained under the stated policy. The production check used a synthetic tenant; unrelated resources from a concurrent test were cleaned separately. This is dated evidence for that cleanup path and does not remove later open account-deletion findings.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenant Unrelated concurrent-test cleanup separate Billing tombstones retained Later deletion findings remain public",
"href": "/evidence/phase-b38"
},
{
"id": "gate-b38-prod",
"title": "Historical account-deletion cleanup correction",
"text": "On 8 September 2026, regression and production-canary evidence reported that deleting a synthetic account removed its active-session resources before destroying its key. The prior behavior left unusable session residue and caused health checks to fail. The corrected deletion refused to report success if teardown failed. Billing-related tombstone records were retained under the stated policy. The production check used a synthetic tenant; unrelated resources from a concurrent test were cleaned separately. This is dated evidence for that cleanup path and does not remove later open account-deletion findings.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenant Unrelated concurrent-test cleanup separate Billing tombstones retained Later deletion findings remain public",
"href": "/evidence/phase-b38"
},
{
"id": "gate-b39",
"title": "Historical checkpoint-timeout correction",
"text": "The 8 September 2026 record reports a correction for checkpoints that exceeded the previous fixed deadline. Tests verified an adjusted time allowance, bounded recovery after a timeout and explicit recording of failed checkpoints. The page reports deployment verification before the later soak resumed. Measurements came from one host and one storage class; the calibration did not track storage that degraded while the service was running. A prematurely started soak on the previous build was stopped, so it is not evidence for the corrected build.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "One host/storage class Calibration not continuously tracking degradation Old-build soak stopped, not proof of fix",
"href": "/evidence/phase-b39"
},
{
"id": "gate-b44",
"title": "Historical production credit-authorization correction",
"text": "On 17 September 2026, laboratory tests reproduced and then rejected unauthorized customer credit changes after the correction. A separate production check with a disposable unfunded account confirmed that customer credentials could not change credit, balances stayed unchanged and legitimate ledger reading still worked. A point-in-time ledger review found no unexpected grant-reference categories in the surviving records. That review is not a guarantee that no misuse ever occurred. Some broader laboratory reruns were deferred under a documented exception; the scope of this production check remains credit authorization.",
"status": "green",
"date": "17 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Disposable unfunded tenant; verification granted nothing Point-in-time ledger review only Broader rerun exception Historical, not current exploitability proof",
"href": "/evidence/phase-b44"
},
{
"id": "gate-b44-prod",
"title": "Historical production credit-authorization correction",
"text": "On 17 September 2026, laboratory tests reproduced and then rejected unauthorized customer credit changes after the correction. A separate production check with a disposable unfunded account confirmed that customer credentials could not change credit, balances stayed unchanged and legitimate ledger reading still worked. A point-in-time ledger review found no unexpected grant-reference categories in the surviving records. That review is not a guarantee that no misuse ever occurred. Some broader laboratory reruns were deferred under a documented exception; the scope of this production check remains credit authorization.",
"status": "green",
"date": "17 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Disposable unfunded tenant; verification granted nothing Point-in-time ledger review only Broader rerun exception Historical, not current exploitability proof",
"href": "/evidence/phase-b44"
},
{
"id": "gate-b48",
"title": "Historical prepaid-usage settlement correction",
"text": "On 17 September 2026, laboratory tests reproduced a defect where session usage did not consume prepaid credit, then recorded correct one-time settlement, charges only for additional usage, parking at exhaustion and resume after a top-up. Production evidence later that day recorded outstanding usage debited once and the funds restriction triggered. No affected session was running, so production parking at exhaustion remained untested. An amendment on 18 September corrected the monitoring count without changing the settlement verdict. The record does not disclose account-linked balances or prove current deployment behavior.",
"status": "green",
"date": "17 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Production running-session parking remained untested Original monitoring count corrected on 18 September Laboratory parking used a simulated service",
"href": "/evidence/phase-b48"
},
{
"id": "gate-b48-prod",
"title": "Historical prepaid-usage settlement correction",
"text": "On 17 September 2026, laboratory tests reproduced a defect where session usage did not consume prepaid credit, then recorded correct one-time settlement, charges only for additional usage, parking at exhaustion and resume after a top-up. Production evidence later that day recorded outstanding usage debited once and the funds restriction triggered. No affected session was running, so production parking at exhaustion remained untested. An amendment on 18 September corrected the monitoring count without changing the settlement verdict. The record does not disclose account-linked balances or prove current deployment behavior.",
"status": "green",
"date": "17 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Production running-session parking remained untested Original monitoring count corrected on 18 September Laboratory parking used a simulated service",
"href": "/evidence/phase-b48"
},
{
"id": "gate-b51",
"title": "Historical signup-credit unit correction",
"text": "On 18 September 2026, isolated tests reproduced a wrong-unit configuration that delivered less signup credit than advertised, then passed after the correction. The new checks rejected the mistaken unit and used the approved price-book value when encountering the stored error. The production section reports deployment and remediation of the undercredited account; repeating the remediation did not add credit again. The automated assertions used an isolated test identity provider with no live fleet or model calls. Production remediation is a separate recorded observation, not an independently repeated audit test.",
"status": "green",
"date": "18 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Isolated automated test Production remediation separately reported No live fleet/model calls in gate",
"href": "/evidence/phase-b51"
},
{
"id": "gate-b52",
"title": "Historical stale-price-override correction",
"text": "On 18 September 2026, isolated tests verified that an old credit-setting override stopped governing after the approved price book changed, with the reason made visible. An explicitly reconfirmed setting could then apply under the new version. The record reports production deployment, but states that no override was active, so the deployment itself did not exercise removal of a stale override. The functional proof came from an isolated test identity provider without live fleet or model calls. This record does not define today’s signup-credit amount.",
"status": "green",
"date": "18 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "No active override during production deploy Functional proof isolated No current credit-price promise",
"href": "/evidence/phase-b52"
},
{
"id": "gate-b53",
"title": "Historical missing-signup-credit correction",
"text": "The 18 September 2026 record discloses that some Google signups did not receive the credit advertised on the website. Isolated tests reproduced the failure and then verified that eligible signups received the approved credit even when their identity provider supplied no account-age information. The production section reports deployment, correction of affected Google accounts and clearing of stale low-credit alerts. The automated tests did not use live identity providers or the fleet. This record preserves the customer impact and reported remediation without publishing account identities or balances.",
"status": "green",
"date": "18 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Some Google signups did not receive the credit advertised on the website. Automated proof isolated Production remediation separately recorded",
"href": "/evidence/phase-b53"
},
{
"id": "gate-billing-ledger-prod",
"title": "Customer-facing hosted service checks",
"text": "On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.",
"href": "/evidence/phase-hosted-prod"
},
{
"id": "gate-browser-console-prod",
"title": "Customer-facing hosted service checks",
"text": "On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.",
"href": "/evidence/phase-hosted-prod"
},
{
"id": "gate-byok-prod",
"title": "Customer-facing hosted service checks",
"text": "On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.",
"href": "/evidence/phase-hosted-prod"
},
{
"id": "gate-ckpt-resume-prod",
"title": "Historical production checkpoint-and-resume canary",
"text": "On 19 September 2026, a synthetic tenant used the production customer routes to create, checkpoint, stop and resume one session. Its memory-backed witness persisted and its counter continued; the completed run cleaned up its test resources. This used one image, one guest size and one resume, with synthetic paid status. It tested the checkpoint route’s orderly stop, not a hard machine kill. Two earlier aborted attempts leaked sessions and were recorded as an open account-deletion issue; the successful cleanup does not erase that limitation.",
"status": "green",
"date": "19 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "One session/image/size/resume Synthetic paid status Orderly stop, not hard kill Two aborted attempts leaked sessions",
"href": "/evidence/phase-ckpt-resume-prod"
},
{
"id": "gate-cli-1",
"title": "Historical public-client installation checks",
"text": "Checks recorded on 1 September 2026 verified installation of the released client, successful sign-in and session start on a fresh Linux virtual machine, and rejection of altered installer/update downloads. macOS checks used a cleaned environment rather than a fresh operating-system installation. A 2 September rerun after the repository transfer verified the new distribution paths and compatibility redirects. Windows installation was not covered. These results apply to those historical client releases and environments, not all current installers or supported platforms.",
"status": "green",
"date": "1 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "macOS environment not fresh VM Windows not tested Historical release paths",
"href": "/evidence/phase-cli-1"
},
{
"id": "gate-cp-1",
"title": "Historical account persistence and callback tests",
"text": "On 30 August 2026, laboratory checks verified that an account and its session survived a service restart, deletion removed identifying account data while retaining a tombstone, and the deleted account’s session was refused. Altered sign-in callback state created no account. Session-cookie protections were also checked. The login round trip used a test identity provider; Google and GitHub redirect construction was checked without contacting those providers. This is not a live-provider authentication test or proof of document-specific authorization.",
"status": "green",
"date": "30 August 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Test identity provider Google/GitHub wiring structural only No scoped-document authorization test",
"href": "/evidence/phase-cp-1"
},
{
"id": "gate-cp-2",
"title": "Historical device-login and token-revocation tests",
"text": "On 30 August 2026, laboratory tests completed the client’s device-login flow, bound the resulting token to the approving account and verified that revocation took effect on the next request. Expired and mismatched approval codes were rejected. The inspected logs and database did not contain the fixture token; it appeared only in the tested client credential store. This exercised the Linux storage path. The macOS keychain path was not tested here, and the result is not a general claim that no token can ever leak.",
"status": "green",
"date": "30 August 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Linux credential store tested macOS keychain untested here Fixture inspection does not prove universal absence of secrets",
"href": "/evidence/phase-cp-2"
},
{
"id": "gate-cp-3",
"title": "Historical onboarding and key-vault tests",
"text": "On 30 August 2026, laboratory checks verified that onboarding state survived a restart, stored a fixture provider key in encrypted form and returned only a masked key reference. Cross-account attempts to read or change the vault were refused. The service also refused an unsafe key-storage configuration. The fixture key was deliberately not a real provider credential, and no live provider call was made. The evidence applies to those inspected responses and stores; it does not certify every current production secret-handling path.",
"status": "green",
"date": "30 August 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Non-provider fixture key No live provider call Inspected paths only",
"href": "/evidence/phase-cp-3"
},
{
"id": "gate-cp-c1",
"title": "Historical usage-meter reconciliation test",
"text": "On 31 August 2026, a test on hosted hardware reconciled a usage statement against its token, session-time and storage records. The total matched its priced lines, an altered meter record failed reconciliation, and an unreadable source was shown as unavailable rather than zero. A real billed model call supplied the token measurement. This test did not exercise card handling or checkout. It verifies the recorded statement and reconciliation checks, not today’s prices or every future bill.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "No checkout/card path Missing source is unavailable, not zero Historical statement, not current prices",
"href": "/evidence/phase-cp-c1"
},
{
"id": "gate-cp-c2",
"title": "Historical credit-ledger and funds-interlock test",
"text": "On 31 August 2026, a hosted-hardware test with a real virtual-machine session verified that repeated credit operations did not duplicate a debit and that exhaustion checkpointed and parked the session, which later resumed after a top-up. A 17 September amendment changed the test because customers were no longer allowed to invoke credit-changing operations. The amended ledger checks run in isolation, with settlement supported by separate evidence. The original assertions describe the earlier interface; they are not proof that those customer operations remain available or that the amended test reran in production.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Original interface superseded Amended ledger proof isolated No production rerun inferred from amendment",
"href": "/evidence/phase-cp-c2"
},
{
"id": "gate-cp-c3",
"title": "Historical payment-notification verification test",
"text": "On 31 August 2026, a laboratory sandbox test verified that an authenticated payment notification added credit once, repeated delivery or another notification for the same order added nothing further, and a forged notification was refused. The simulated payment, order and credit records reconciled. A mock payment service and test signing secret were used; this was not a live card purchase or a test of the hosted checkout page. KeepState’s usage statement is separate from the payment provider’s tax receipt. Later live-payment assurance requires its own evidence.",
"status": "green",
"date": "31 August 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Mock provider and test secret No live card purchase No hosted checkout-page test Usage statement not tax receipt",
"href": "/evidence/phase-cp-c3"
},
{
"id": "gate-cp-c4",
"title": "Historical usage-statement integrity test",
"text": "On 31 August 2026, a test on hosted hardware verified that a downloadable usage statement matched its source statement and priced lines, retained a verifiable relationship to the prior statement and detected alteration. It was explicitly a usage statement, not a tax receipt. A 17 September amendment changed how the test account was funded after customer credit-changing operations were closed. This version says the amended rerun was still pending an authorized test window. The original passing result does not establish a completed amended production rerun.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Usage statement, not tax receipt Amended rerun pending in this version Original result not amended production proof",
"href": "/evidence/phase-cp-c4"
},
{
"id": "gate-cp-c5",
"title": "Historical low-credit notification test",
"text": "On 31 August 2026, laboratory tests verified that low-credit alerts were calculated from server-side account records, appeared once in the console and were not falsely marked emailed when email was unavailable. With a mock delivery endpoint, an alert reached that endpoint before being marked sent. Live authenticated email was outside this test. A 17 September amendment changed test funding; its attempted rerun was not held because the laboratory environment was unreachable. The original result remains historical, and the aborted attempt is not a passing rerun.",
"status": "green",
"date": "31 August 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Mock email endpoint Live authenticated email separate Amended rerun trial not held",
"href": "/evidence/phase-cp-c5"
},
{
"id": "gate-cp-c6",
"title": "Audited activation and welcome-screen behavior",
"text": "On 31 August 2026, a laboratory test recorded three passing assertions: accounts could not activate themselves, authorized activation was audited, and queued accounts received no start instructions while active accounts received usable welcome instructions. A negative control confirmed that the activation restriction mattered. The welcome test checked instruction structure, not the production destination. This historical result does not establish that the same activation policy or interface is deployed today.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Laboratory run; production URL not verified. Three per-assertion passes, not a product-wide verdict.",
"href": "/evidence/phase-cp-c6"
},
{
"id": "gate-cr-1",
"title": "Approve the task and its checks before running",
"text": "On 16 September 2026, four laboratory assertions passed for Cruise preparation. The client drafted a task manifest, required explicit approval, and refused execution if the approved task or tests changed. A repository without discoverable tests was rejected. Preparation and approval made no service requests in the test. Discovery covered pytest and Go tests only; other test systems were not established by this record. This is a dated client-contract result, not current production verification.",
"status": "green",
"date": "16 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Only pytest and Go test discovery covered. Laboratory client and recording service; no live execution proved.",
"href": "/evidence/phase-cr-1"
},
{
"id": "gate-cr-2",
"title": "Retry from the saved state within the approved spending limit",
"text": "On 16 September 2026, a laboratory fixture completed a three-step job after two failed attempts. Each retry started from the same saved state. The test refused an unsupported model choice, stopped when the remaining budget could not cover another attempt, and made no retry after the final attempt. A negative control detected a disabled budget check. These are fixture results, not measured production costs or a savings guarantee.",
"status": "green",
"date": "16 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Laboratory fixture only No production spending or savings guarantee",
"href": "/evidence/phase-cr-2"
},
{
"id": "gate-cr-3",
"title": "Each attempt starts from the same saved workspace",
"text": "On 17 September 2026, six infrastructure assertions passed for isolated attempts from one saved workspace, restored content, periodic save points and cleanup on hosted infrastructure. The addendum records that every attempt failed: the test accounts lacked the required provider key, so no inference call or task acceptance occurred. Successful-attempt parking was conditional and was not demonstrated. This one-job run supports the tested infrastructure behavior, not task success or acceptance rates. Checkpoint pauses remain host-specific measurements.",
"status": "green",
"date": "17 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "No inference call or task acceptance: test accounts had no provider key. Successful-attempt parking was conditional and not demonstrated. One job; checkpoint pauses are host-specific.",
"href": "/evidence/phase-cr-3"
},
{
"id": "gate-cr-4",
"title": "Independent verification rejects a forged pass",
"text": "On 16 September 2026, a hosted-service test recorded four passing checks: the candidate was extracted and the child stopped before verification; the verifier ran on the host using approved tests; verdicts carried signed test references; and a forged pass log was rejected. A negative control showed that compromised verifier independence could invalidate the result. The tested job ended in review after its single attempt failed. This establishes the tested verification boundary, not universal task success or current deployment state.",
"status": "green",
"date": "16 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "One hosted attempt The job ended in review, not acceptance Current deployment independently unverified",
"href": "/evidence/phase-cr-4"
},
{
"id": "gate-cr-5",
"title": "Unresolved attempts remain visible in the job record",
"text": "On 16 September 2026, four laboratory assertions passed for job accounting. Attempts progressed through their recorded lifecycle, reported spend matched reconciled attempts, and the event history passed its integrity check. When usage remained unresolved, the attempt stayed visible and the job could not be accepted. A negative control caught an attempted acceptance despite that ambiguity. These checks used a fixture service with a simulated reconciliation timeout; they did not measure a live provider’s reporting delay.",
"status": "green",
"date": "16 September 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Fixture backend and modeled timeout. No live-provider reconciliation-latency claim.",
"href": "/evidence/phase-cr-5"
},
{
"id": "gate-cr-6",
"title": "Job status and recorded results",
"text": "The original laboratory record is dated 16 September 2026. A 17 September amendment changed test funding after customer credit grants were disallowed; the revised normal and negative-control runs each passed eight checks. They rejected a changed approval digest and a second concurrent free-tier job, verified event history, prevented acceptance of unresolved attempts and resume outside review, and displayed missing spend as unavailable. Customers could not grant themselves credit. These local tests used a test identity service and do not establish current production limits or availability.",
"status": "green",
"date": "16 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Laboratory services and test identity provider. Customer credit changes require authorized support. Historical tier constraint requires current product confirmation.",
"href": "/evidence/phase-cr-6"
},
{
"id": "gate-cr-7",
"title": "Offline help and verified artifact downloads",
"text": "On 16 September 2026, five laboratory assertions passed for the Cruise client. Twenty help invocations made no service requests, status and logs preserved the recorded states and event order, and a modified artifact was refused before writing an output file. Workspaces larger than 50 MiB were rejected before upload; the measured size excluded Git and KeepState metadata directories. This record covers the tested client contract and recording service, not every current client release.",
"status": "green",
"date": "16 September 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "50 MiB threshold and exclusions belong to this tested release. Recording service, not full production integration.",
"href": "/evidence/phase-cr-7"
},
{
"id": "gate-cr-8-prod",
"title": "A hosted job through the customer-facing service",
"text": "On 17 September 2026, seven assertions passed in a production-hosted check with a bring-your-own-key account. A job was accepted, an unrelated test tenant could not see it, its event history verified, and the downloaded artifact matched its recorded digest. A successor run that day repeated the checks after client corrections. Monetary spend remained unavailable: usage records did not include money, so zero was not a valid cost claim. Acceptance was an observed outcome, not a required success rate. Artifact verification was conditional on acceptance, and the record does not establish other workloads or later deployments.",
"status": "green",
"date": "17 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Monetary spend unavailable; usage does not establish money spent. Limited job/account scenario; acceptance is not a performance promise. Artifact verification applies only to accepted runs.",
"href": "/evidence/phase-cr-8-prod"
},
{
"id": "gate-cr-9",
"title": "Review, resume and cancellation after failed attempts",
"text": "On 16 September 2026, four fixture-based laboratory assertions passed. Exhausting the approved attempt plan moved a job into review with a notification event. Extending the plan from review allowed the tested job to complete; cancellation released the retained workspace. The final permitted attempt was not silently retried. Resume was allowed only from review, and the retained parent remained the customer’s recovery option until release. This record does not prove live notification delivery or production availability.",
"status": "green",
"date": "16 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Fixture backend. Resume only from review. Notification event is not proof of message delivery.",
"href": "/evidence/phase-cr-9"
},
{
"id": "gate-em-1",
"title": "Email is recorded before delivery is claimed",
"text": "The original record dated 2 September 2026 reports three passing assertions across laboratory and hosted email checks. Simulated interruptions left mail queued, and sent status required provider acceptance. The provider verified the domain’s email authentication configuration and accepted a welcome message. The interruption tests used a mock provider; provider acceptance does not prove inbox delivery or receiver authentication headers. A 17 September amendment changed test funding, but its rerun was not held because the laboratory host was unreachable. The amendment therefore adds no new passing run.",
"status": "green",
"date": "2 September 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Mock provider for interruption tests. Provider acceptance does not prove recipient receipt or receiver headers. 17 September amended rerun was not held because the laboratory host was unreachable.",
"href": "/evidence/phase-em-1"
},
{
"id": "gate-email-prod",
"title": "Production transactional email sending check",
"text": "On 9 September 2026, a production transactional-email check recorded a green result. The provider reported the sending domain authenticated, a DMARC policy was present in public DNS, and the deployed service’s message was accepted by the mail provider. The check did not independently derive every DNS authentication record or measure inbox placement. It did not establish delivery to a subscriber list; subscriber fan-out and an unsubscribe route were absent in this record.",
"status": "green",
"date": "9 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Provider authentication-state reporting, not independent DNS derivation. No inbox-placement or subscriber-delivery proof. Subscriber fan-out and unsubscribe route absent.",
"href": "/evidence/phase-email-prod"
},
{
"id": "gate-er-1",
"title": "Account erasure and recovery isolation",
"text": "On 7 September 2026, six recorded assertions passed for account-level erasure on hosted infrastructure. Deleted-account checkpoints became unreadable after key destruction, including from the tested backup copy, while another account’s matching content remained restorable. Identity was removed from retained billing records. The rollout used team/test data; a cleanup error also removed the test canary’s chunks, with no reported customer impact. A fresh post-cleanup checkpoint subsequently resumed. This record does not establish that every backup or future deployment follows the same key-management conditions.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Team/test rollout only; cleanup error and no reported customer impact retained. Backup recovery depends on stated key-management conditions.",
"href": "/evidence/phase-er-1"
},
{
"id": "gate-er-1c",
"title": "Deletion fails closed when erasure cannot be confirmed",
"text": "On 7 September 2026, four assertions passed using production binaries with disposable data. The customer deletion path removed credentials and made checkpoint data unrecoverable through the tested backup-recovery path. It did not report deletion if key destruction could not be confirmed. A second tenant remained restorable. Production services were not changed by this test; it exercised their deletion code in isolation. Recovery protection depends on keys being excluded from the restored backup.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Disposable data; production daemons untouched. Backup/key separation is required.",
"href": "/evidence/phase-er-1c"
},
{
"id": "gate-er-1c-prod",
"title": "Deletion verified on the running hosted service",
"text": "On 8 September 2026, three production-hosted assertions passed for deletion of a synthetic tenant. The customer-facing deletion operation destroyed checkpoint access, removed stored provider credentials and removed the account identity without affecting another tenant. A copied checkpoint was unrecoverable under the stated key-management conditions. The fix was deployed during this work; the result applies from that deployment, not earlier builds. The copy was made for the test because no scheduled checkpoint-store backup existed.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenant. Attests only from the corrected deployment onward. Test-made backup copy; no scheduled checkpoint-store backup.",
"href": "/evidence/phase-er-1c-prod"
},
{
"id": "gate-er-1d",
"title": "Deleted credentials stay unreadable in the tested backup",
"text": "The credential-erasure record dated 8 September 2026 reports three passing hosted test assertions. A database backup recovered the test credential before account deletion but could not afterward, even with the master key retained. A negative control confirmed that destroying the account key was necessary. The base test required older entries to be rewritten; an appended production section dated 7 September reports migrating and verifying all four existing entries. Erasure protection depends on keeping account keys outside the database backup set. The record does not establish later deployments or independent copies of those keys.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Separate account-key custody and exclusion from the database backup set are required. Base test and production migration appendix are distinct evidence. Later deployments and independent key copies are outside this record.",
"href": "/evidence/phase-er-1d"
},
{
"id": "gate-er-1d-prod",
"title": "Credential erasure protection enabled in production",
"text": "The production record dated 8 September 2026 reports enablement on 7 September. Its single assertion passed: all four existing credential entries were migrated to per-account protection and verified to require their account keys rather than the master key alone; a live model request still succeeded. This supports the reported migration at that time. Credential erasure depends on keeping account keys outside the database backup set. The record does not certify later entries, later deployments or independent copies of key material.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Erasure depends on account keys being excluded from database backups. Historical migration result, not continuing certification. 8 September record date differs from 7 September reported enablement.",
"href": "/evidence/phase-er-1d-prod"
},
{
"id": "gate-fo-1",
"title": "Resume on another host after an interruption",
"text": "On 5 September 2026, four hosted assertions passed for moving saved sessions to another host after an isolated service failure. The tested sessions preserved in-memory state, each move produced one event and a customer notice, and duplicate resume attempts were prevented. The test used shared durable storage within one region. Work since the latest checkpoint could be lost. Billing reconciliation on the replacement host was still pending in this record; these results do not prove recovery from every host or regional failure.",
"status": "green",
"date": "5 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Isolated service termination, same-region shared storage. Checkpoint-bounded loss window. Replacement-host billing reconciliation pending in this record.",
"href": "/evidence/phase-fo-1"
},
{
"id": "gate-fo-1b",
"title": "Billing state across a real host loss",
"text": "The record combines a hosted failover rehearsal dated 6 September 2026 and production enablement dated 7 September. Six assertions report billing continuity after real host deallocation: recorded spend survived, an identical retry was not billed again, and budget enforcement used the carried balance. Settlement stayed blocked until reconciliation passed. One run measured 47 seconds from deallocation to reconciliation; this is not an availability promise. The enablement addendum records a live production check. Session recovery relied on a separately tested mechanism, and regional disaster recovery was not established.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "47 seconds is one run’s deallocation-to-reconciliation measurement. Shared storage and separately tested guest recovery are dependencies. Settlement remains blocked until reconciliation completes. Production enablement is a separate 7 September addendum.",
"href": "/evidence/phase-fo-1b"
},
{
"id": "gate-fork-prod",
"title": "Customer-facing hosted service checks",
"text": "On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.",
"href": "/evidence/phase-hosted-prod"
},
{
"id": "gate-gate-0",
"title": "Historical environment-readiness test",
"text": "On 28 July 2026, the laboratory environment passed readiness checks for running hardware-virtualized guests and verifying its pinned software. A deliberately altered integrity value was rejected. The result supports that environment’s readiness and integrity checks at that point in time. It does not demonstrate an agent workload, hosted availability or ongoing production health.",
"status": "green",
"date": "28 July 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Point-in-time laboratory hygiene Single altered integrity value No hosted availability claim",
"href": "/evidence/phase-0"
},
{
"id": "gate-gate-1",
"title": "Historical memory-continuity test",
"text": "On 28 July 2026, a laboratory guest resumed after its virtual-machine process was forcibly stopped. Its in-memory witness remained the same and its counter continued advancing; a fresh boot failed the continuity checks. The guest used a read-only filesystem, so the witness could not be reconstructed from a saved disk file. This was a small counter guest, not an AI agent or a production workload, and establishes only the recorded memory-continuity result.",
"status": "green",
"date": "28 July 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Counter guest, not AI agent Read-only filesystem test Laboratory, not production",
"href": "/evidence/phase-1"
},
{
"id": "gate-gate-2a",
"title": "Historical session lifecycle test",
"text": "On 28 July 2026, laboratory checks ran a session through checkpoint, forced stop and resume while preserving its in-memory witness and continuing counter. Repeating the test confirmed the lifecycle could run again, and substituting a fresh start failed the continuity checks. This test did not verify clock resynchronization; a separate general-image demonstration was reported as data, not an asserted result. The evidence concerns the recorded laboratory lifecycle, not current hosted availability.",
"status": "green",
"date": "28 July 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Clock resynchronization untested General-image smoke result is data only Laboratory scope",
"href": "/evidence/phase-2a"
},
{
"id": "gate-gate-2b",
"title": "Historical checkpoint portability and integrity test",
"text": "On 28 July 2026, a laboratory test restored a session from its stored checkpoint on a second service instance, created branches with recorded ancestry, and rejected corrupted checkpoint data. An interrupted checkpoint publication left the prior checkpoint intact. Both service instances ran on the same laboratory host with compatible software and local storage; this was not a cross-host disaster-recovery test. Branch differences came from staggered execution, not agent steering.",
"status": "green",
"date": "28 July 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Second instance on same host Compatible software/local storage Not cross-host recovery Branch divergence not steering",
"href": "/evidence/phase-2b"
},
{
"id": "gate-gate-3",
"title": "Historical interrupted-call replay test",
"text": "On 28 July 2026, a laboratory Claude Code session survived a forced stop and resume. Its records reconciled, and a completed interrupted response was replayed without being billed again. The test also detected deliberately degraded replay and record behavior. The result depends on the client returning for the held response; it is not a promise about every agent’s retry behavior. The specimen’s task outcome was reported separately and did not determine whether the infrastructure test passed.",
"status": "green",
"date": "28 July 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Claude Code client scope Client retry behavior separate Task outcome not asserted Some billing checks alone remain true in degraded system",
"href": "/evidence/phase-3"
},
{
"id": "gate-gate-4",
"title": "Historical unattended recovery demonstration",
"text": "On 29 July 2026, a laboratory demonstration automatically checkpointed an agent, stopped both its virtual machine and supervising service, and resumed the session. The recorded run passed its recovery and cleanup checks. A run without a usable checkpoint correctly failed. The asserted endpoint was the agent being alive after resume; completion of its task was separate reported data. Reported component timings were measurements, while monetary conversions were estimates. This is a single demonstration result, not a general recovery-time or task-success guarantee.",
"status": "green",
"date": "29 July 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Alive after resume, not task completion Component timings reported, not all asserted Dollar figures estimated Single laboratory demonstration",
"href": "/evidence/phase-4"
},
{
"id": "gate-gate-5",
"title": "Historical branch-isolation and steering test",
"text": "On 29 July 2026, laboratory checks created branches from a shared checkpoint, applied distinct steering inputs and verified separate branch histories and filesystems. A test-written marker in one branch did not appear in the others. Identical steering inputs were correctly detected as failing the distinctness requirement. The marker was written by the test harness, not the agent: it proves filesystem isolation, not that an agent obeyed its instructions. Agent behavior remains measured data rather than a guaranteed result.",
"status": "green",
"date": "29 July 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Marker written by test harness Isolation is not agent obedience Laboratory scope",
"href": "/evidence/phase-5"
},
{
"id": "gate-gate-6",
"title": "Historical long-running checkpoint reliability test",
"text": "The snapshot records this historical soak as not required for the release rerun; its old result is not a fresh run. The record sealed on 31 July 2026 reports a replacement seven-hour laboratory soak with successful scheduled checkpoints and restoration of an older checkpoint. An earlier run exposed a defect that stopped repeated checkpointing; the record reports its correction and retest. It also discloses that cleanup destroyed the originally preserved checkpoint, so its restoration was not performed; the later restoration used a replacement artifact. One workload and one soak do not establish a general reliability rate. The separate 25-cycle sample supports only its stated confidence interval, not guaranteed success.",
"status": "not-required",
"date": "31 July 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "One soak/one workload Original preserved artifact destroyed; original restoration not performed Replacement restoration is distinct 25-cycle confidence limits remain visible",
"href": "/evidence/phase-6"
},
{
"id": "gate-gate-7",
"title": "Historical second-agent compatibility test",
"text": "The original record of 31 July 2026 and rerun of 30 August tested a second agent, aider, and a fresh laboratory installation. The gateway retained the interrupted response in all four reported runs; aider retried successfully in three. That difference is a material client limitation. The second installation used another virtual machine on the same physical computer. A later attempt was not counted as a completed trial because its environment was not ready. These results support the named agent and setup, not compatibility with every agent or guaranteed resumption.",
"status": "green",
"date": "31 July 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Aider retry 3/4 versus held response 4/4 Second VM on same physical machine Trial-not-held excluded explicitly Not universal compatibility",
"href": "/evidence/phase-7"
},
{
"id": "gate-hosted-attach-prod",
"title": "Customer-facing hosted service checks",
"text": "On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.",
"href": "/evidence/phase-hosted-prod"
},
{
"id": "gate-hosted-prod",
"title": "Customer-facing hosted service checks",
"text": "On 8 September 2026, six production-hosted assertions passed for synthetic tenants. The check covered authenticated account data, onboarding redirection, write-only credential submission, an authorized session fork, attachment negotiation and billing displays. An unrelated real tenant remained unchanged. Limits matter: the credential was synthetic, the paid state was simulated, attachment covered the handshake rather than an interactive session, and the billing account had no spend. The rendered console and larger fork fan-outs were not verified.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Synthetic tenants, key and paid-state setup. No rendered-console or interactive-keystroke proof. No-spend billing case; larger fan-outs untested.",
"href": "/evidence/phase-hosted-prod"
},
{
"id": "gate-hr-1-codex",
"title": "Separate recovery, metering and retry support checks",
"text": "On 5 September 2026, eight hosted assertions passed across two supported provider paths. Test sessions preserved in-memory state through save, interruption and resume; real provider calls were metered; identical retries avoided a second charge; missing credential access failed closed. The tests used a base guest exercising the integration configuration, rather than proving every behavior of a full agent application. Whether a request is retried remains the client’s decision. These results apply to the tested combinations.",
"status": "green",
"date": "5 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Base guest/configuration surface, not every native agent workflow. Retry issuance is client-controlled.",
"href": "/evidence/phase-hr1"
},
{
"id": "gate-hr-1-openrouter-generic",
"title": "Separate recovery, metering and retry support checks",
"text": "On 5 September 2026, eight hosted assertions passed across two supported provider paths. Test sessions preserved in-memory state through save, interruption and resume; real provider calls were metered; identical retries avoided a second charge; missing credential access failed closed. The tests used a base guest exercising the integration configuration, rather than proving every behavior of a full agent application. Whether a request is retried remains the client’s decision. These results apply to the tested combinations.",
"status": "green",
"date": "5 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Base guest/configuration surface, not every native agent workflow. Retry issuance is client-controlled.",
"href": "/evidence/phase-hr1"
},
{
"id": "gate-kill-guard",
"title": "Protect unsaved work from accidental termination",
"text": "On 31 August 2026, three hosted assertions passed for the termination guard. An ordinary termination request with unsaved work was refused and the session stayed alive, with safer choices explained. Explicit forced termination proceeded; a session with a saved checkpoint could be terminated normally. The protection was enforced by the service as well as the client. This guard protects unsaved work, not all destructive actions, and an explicit force request can still discard that work.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Guard concerns unsaved work. Explicit force permits destruction.",
"href": "/evidence/phase-kill-guard"
},
{
"id": "gate-l-1",
"title": "Historical accepted-work test across two provider families",
"text": "A confirmation run on 6 September 2026 completed one bounded coding fixture through each of two live provider families. The initial attempts failed; later attempts produced accepted artifacts after independent checks. Seven checks passed, including complete-job accounting. Publication was limited to a scoped authorization: no external pull request was created. A negative control showed that verifier independence mattered. The sample is two task arms with deliberately constrained initial attempts, not a general task-success rate, comparative model ranking or cost-savings claim.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "One bounded fixture in two provider families Initial attempts were deliberately resource-constrained No external publication or general success rate",
"href": "/evidence/phase-l-1"
},
{
"id": "gate-lr-1",
"title": "Rehearsal of onboarding, usage and billing reconciliation",
"text": "On 31 August 2026, a hosted rehearsal passed all ten steps in the recorded customer journey and two summary assertions. It covered activation, a supplied provider key, crediting, model use, pause/resume, statement reconciliation and a refund returning the balance to zero. Revoked access and a modified statement were rejected. The identity provider and payment service were test systems: this was a rehearsal, not proof that a real new customer completed live checkout or received a real refund.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Test identity provider and mock payment sandbox. No live customer checkout or live refund established.",
"href": "/evidence/phase-lr-1"
},
{
"id": "gate-pv-1",
"title": "Metering and replay on an additional provider path",
"text": "On 4 September 2026, six hosted assertions passed for an additional compatible provider path. Deterministic tests checked credential routing, usage accounting, ordered records and replay of an identical request without another charge. A separate live provider call also succeeded and its identical retry was replayed. The live gateway ran in isolation and did not prove the full agent-plus-session-interruption combination. The replay guarantee concerns identical retries; it does not mean all similar requests are free.",
"status": "green",
"date": "4 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Mechanics used a deterministic mock. Live gateway was isolated; full agent recovery separate. Replay applies to identical requests.",
"href": "/evidence/phase-pv-1"
},
{
"id": "gate-pv-2",
"title": "Change models only while a session is saved",
"text": "On 4 September 2026, three hosted test assertions passed for changing a session’s model. A change was refused while the session was running, accepted while it was checkpointed, and used by the next request without resetting the request lineage. These tests used a simulated compatible provider. They covered compatible clients through the tested routing path; native clients for the other protocol were explicitly outside this result. Live continuation quality was not measured here.",
"status": "green",
"date": "4 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Mock upstream. Native clients for the other protocol not claimed. No live continuation-quality measurement.",
"href": "/evidence/phase-pv-2"
},
{
"id": "gate-pv-3",
"title": "Continue a fixture across provider protocols",
"text": "On 7 September 2026, four hosted assertions passed for continuing a tested task across provider protocols at a checkpoint boundary. The live continuation included tool use and an independently verified artifact, while lineage and metering remained continuous. A failed continuation was refused and the prior workspace restored for fallback. Deterministic translation checks were separate from the live run. The result is fixture-specific and model-dependent; streaming translation was unsupported and explicitly refused in this version.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Fixture-specific continuation; quality depends on model. Streaming unsupported in this version. Deterministic and live checks have separate scopes.",
"href": "/evidence/phase-pv-3"
},
{
"id": "gate-pv-3-prod",
"title": "Production check of cross-provider translation",
"text": "On 8 September 2026, one production-hosted assertion passed for a real translated request on the team’s account. The request reached a live compatible model, was recorded and metered, and the service stayed healthy. This was a one-call production check, not a broad task-success study. Streaming responses were buffered rather than incremental, and replayed thinking blocks were omitted during translation. Those limitations remain part of the result; current behavior requires separate deployment verification.",
"status": "green",
"date": "8 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "One call on a team account. Buffered streaming; replayed thinking blocks omitted.",
"href": "/evidence/phase-pv-3-prod"
},
{
"id": "gate-pv-3b",
"title": "Native coding-agent continuation across providers",
"text": "On 7 September 2026, three hosted assertions passed with the actual tested coding-agent application and real tool use. A fixture task continued after a checkpoint-boundary provider change and produced an independently verified artifact. Lineage and metering remained continuous; a disabled continuation path failed closed and fallback completed after restoring the workspace. The agent ran in an isolated copy of its environment, with transport tested separately. Quality remains model-dependent, and streaming was buffered rather than incremental.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Isolated application environment; transport proved separately. Fixture and model-dependent result. Buffered, nonincremental streaming.",
"href": "/evidence/phase-pv-3b"
},
{
"id": "gate-quickstart-clean-room",
"title": "Independent quickstart run for release v0.1.3",
"text": "On 25 August 2026, a clean-environment run of the v0.1.3 quickstart completed successfully on an Apple Silicon laboratory machine. The source instructions were hash-checked before execution. The first attempt stopped because a stale checkout existed; a fresh second attempt exited successfully in 176 seconds, preserved session state across resume and completed cleanup. This was a quickstart demonstration, not certification of the separate blocked acceptance test. That broader test remained blocked on two assertions awaiting a human decision.",
"status": "green",
"date": "25 August 2026",
"scope": "Laboratory or bench evidence; production behavior is not inferred from it.",
"limitation": "Two attempts; stale-checkout refusal retained. 176 seconds belongs to one successful attempt. Broader acceptance gate remained blocked.",
"href": "/evidence/quickstart-run-v0.1.3-redacted"
},
{
"id": "gate-rs-1",
"title": "Measured session resume times",
"text": "On 5 September 2026, a hosted measurement ran 20 warm and 20 cold resumes of a 2 GiB guest in interleaved order. All 40 completed. Warm p95 was 25.09 seconds and cold p95 was 47.46 seconds; medians were 18.08 and 33.36 seconds. The cold result met the 60-second target, while the warm result missed the 5-second target. These measurements describe that guest and environment, not a general latency guarantee or current service level.",
"status": "green",
"date": "5 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Warm p95 missed the target. Single tested guest/environment and 40 runs. Measurement, not service guarantee.",
"href": "/evidence/phase-rs1"
},
{
"id": "gate-s1",
"title": "Deterministic checks for the acceptance contract",
"text": "On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.",
"href": "/evidence/phase-s1"
},
{
"id": "gate-s11",
"title": "Accepted-work checks on a plain Linux stand-in",
"text": "On 7 September 2026, three assertions passed for the accepted-work workflow on a plain Linux customer-environment stand-in. A failed attempt restarted from a saved segment boundary, a later attempt completed, and the artifact was independently checked after the remote round trip. The environment’s limitations appeared in the result and missing limitation declarations were rejected. Plain-host recovery did not rewind memory. This was a tested integration configuration, not proof for every customer host or workload.",
"status": "green",
"date": "7 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Customer-environment stand-in. No RAM rewind on plain hosts; segment restart only. Configuration-specific result.",
"href": "/evidence/phase-s11"
},
{
"id": "gate-s2",
"title": "Deterministic checks for the acceptance contract",
"text": "On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.",
"href": "/evidence/phase-s1"
},
{
"id": "gate-s3",
"title": "Live task completion after a failed attempt",
"text": "On 6 September 2026, four hosted assertions passed across two live provider families. A first-stage candidate could pass independent checks; when a first attempt failed, the workspace was restored and a later stage completed the fixture. Calls were metered, and disabling escalation did not turn failure into acceptance. The tasks were tiny supplied fixtures, and recovery here was at workspace-segment level. Memory restoration was established separately; no acceptance-rate, cost-saving or general workload claim follows.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Tiny supplied fixtures. Workspace-level segment restore; memory recovery separate. No savings or general acceptance-rate claim.",
"href": "/evidence/phase-s3"
},
{
"id": "gate-s4",
"title": "Deterministic checks for the acceptance contract",
"text": "On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.",
"href": "/evidence/phase-s1"
},
{
"id": "gate-s5",
"title": "Deterministic checks for the acceptance contract",
"text": "On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.",
"href": "/evidence/phase-s1"
},
{
"id": "gate-s6",
"title": "Deterministic checks for the acceptance contract",
"text": "On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.",
"href": "/evidence/phase-s1"
},
{
"id": "gate-s7",
"title": "Deterministic checks for the acceptance contract",
"text": "On 6 September 2026, six deterministic assertions passed for the acceptance layer. Approved task conditions could not change silently; verification was separated from agent-controlled material; publication was limited to an authorized review branch or pull request; and unresolved attempts blocked settlement. Complete-job accounting matched an independent calculation, while unreadable inputs were shown as unavailable. These were isolated deterministic checks. A live multi-model task and real external publication were separate matters; this record does not establish either.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Deterministic checks, separate from live model execution. Publication authorization limited to review branch or pull request. No external publication proved.",
"href": "/evidence/phase-s1"
},
{
"id": "gate-self-serve-prod",
"title": "Self-service signup: status observation and team attestation",
"text": "On 14 September 2026, the public production status reported self-service signup enabled and operational. Separately, the QA team attested that real-provider signups had completed, automatically activated accounts and received the signup credit during the preceding week. These are different evidence types: the status observation proves the setting was enabled; the successful-signup claim is a team attestation, not an independently driven browser test in this record. Current signup behavior requires a fresh end-to-end check.",
"status": "green",
"date": "14 September 2026",
"scope": "Historical record; any production observation is limited to the dated test described below.",
"limitation": "Enabled status does not prove signup completion. Successful signup is team attestation, not a driven canary.",
"href": "/evidence/phase-self-serve-prod"
},
{
"id": "gate-soak-2",
"title": "Reissued six-hour checkpoint evidence",
"text": "The reissued record spans the 31 August 2026 release and 7 September test evidence. It preserves the original hosted segment’s failure to reach 72 checkpoints: only 68 were recorded, although recovery of old state and bounded memory passed. A later run using the corrected driver reached 72 checkpoints across its own six hours and restored the earlier checkpoint. Reissue did not turn the first result into a pass. This is evidence for two identified test segments, not a guarantee for every workload or current release.",
"status": "green",
"date": "31 August 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Original failed segment retained. Later corrected run is separate. Record spans release and run dates.",
"href": "/evidence/phase-soak-2-reissued"
},
{
"id": "gate-soak-3",
"title": "Encrypted checkpoint soak: one checkpoint missed",
"text": "The encrypted soak failed because one expected checkpoint was missed. The later rerun is a separate record; this failed result stays visible. The run completed in September 2026 and was sealed as failed on 8 September. On the encrypted hosted configuration, a six-hour segment restored its hours-old state and stayed within the other measured durability limits, but only 71 of 72 checkpoints completed because one snapshot exceeded its deadline. The record therefore remains failed. A later amendment points to a successful rerun completed on 19 September; that result belongs to the separate rerun and does not erase this defect or its original outcome.",
"status": "red",
"date": "2026-09-08",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "71/72 failure preserved. Run transcript dated 7 September; failed seal dated 8 September. Later supersession is not a revised original pass.",
"href": "/evidence/phase-soak-3"
},
{
"id": "gate-soak-3b",
"title": "Six-hour encrypted checkpoint rerun completed",
"text": "From 18 to 19 September 2026, one hosted six-hour segment completed all 72 checkpoints after the earlier deadline defect was corrected. Six recorded assertions passed, covering the segment duration, checkpoint count, cadence, bounded memory and absence of checkpoint failures, plus a control against pooling unrelated runs. This rerun used one guest size on one host. It does not turn the earlier failed run into a pass or establish reliability for larger guests, other workloads or the current deployment.",
"status": "green",
"date": "19 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "One guest size and one host. Earlier failed run remains failed. Date taken from body because viewer metadata is incomplete.",
"href": "/evidence/phase-soak-3b"
},
{
"id": "gate-t0-c-p",
"title": "Evidence catalogue titles and source records",
"text": "On 6 September 2026, two deterministic publication assertions passed. Evidence titles and summaries were derived from each record’s declared subject, conflicting manually supplied titles were rejected, and records missing required metadata stopped generation. This checks the consistency of the evidence catalogue, not the truth of every underlying product claim. It covers the record-generation side; whether the public website actually rendered the correct titles required a separate website check.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Catalogue consistency, not product verification. Public-site rendering tested separately.",
"href": "/evidence/phase-t0-c-p"
},
{
"id": "gate-t0-p",
"title": "Public claims and their recorded evidence",
"text": "On 6 September 2026, three deterministic publication assertions passed. Claimed results had to reference specific assertions present in the cited records, rather than just naming a test or release. The check reconciled 152 assertions across 42 test records and rejected empty or invalid references. This establishes traceability within the inspected records, not whether each product result is correct or currently deployed. The public website’s rendering was a separate scope.",
"status": "green",
"date": "6 September 2026",
"scope": "Historical evidence; the scope and environment are described in the recorded result.",
"limitation": "Record linkage, not independent product correctness. Website rendering separate.",
"href": "/evidence/phase-t0-p"
},
{
"id": "gate-agents-prod",
"title": "Agent-session production canary on 30 September 2026",
"text": "On 30 September 2026, a production canary created an agent session, bound the account’s own provider key, started its supervised machine and completed an instruction with Claude Code 2.1.251. Eight normal-run checks passed. A separate negative check refused a model call without a bound key. Usage showed a metered call; dollar cost remained unavailable because no approved model rate existed. Deletion stopped the tested machine, while checkpoint content remained subject to retention. An earlier run printed an empty runner label; a corrected final run followed. This one-account demonstration does not validate every agent, adviser, budget, approval or recovery workflow.",
"status": "green",
"date": "30 September 2026; final normal run at 09:16:29 UTC",
"scope": "One team-operated production account using its own provider key and Claude Code 2.1.251.",
"limitation": "One production account, one supported runner version and a bounded instruction. Eight normal checks and one no-key negative check; model reply content was not the success criterion. The registry declared availability for several features; the canary did not independently exercise every declared workflow. Dollar cost was unavailable because no approved rate was recorded. Session deletion stopped the tested machine; checkpoint-content erasure was not tested. The first run’s runner-label display was corrected before the final normal run; its original pass conditions did not depend on that label. Open findings about approval timing, report integrity and unpriced spending are not closed by this canary.",
"href": "/evidence/phase-agents-prod"
}
]
}
],
"currentDeployment": "Not independently verified by this historical evidence.",
"canonical": "/evidence/gates-current"
}
Download this public summary
This derivative has its own digest. It is not the original transcript or a new test. Historical original records are preserved privately.
Download public JSON →SHA-256 76dc85677ca04262423e595ee69729cd27b4d4ca51aa7f17e6a2a6897cee9d8f