Record readiness evidence binding options across landings

[seathatflowsinourveins]

Scope

• Corrects the readiness binding record to recommend a CC decision, cite SLSA Build Provenance v1.2 for consumed inputs, and state the absence of a hosted check of committed readiness.
• Fix head: 837cbdd9bb87f77425dfa8b038bdd83c2f20595c; one commit after f7c68b785d7d622b35bc37557bf630f128b0deed.
• Base commit: ebe6c223f65bf20d04c56858e8669e0bd76ef3aa.
• Lane: lane:foundation.
• Owned paths touched: docs/decisions/2026-10-09-readiness-evidence-binding.md.

SOTA sources

• Command-center authority at ebe6c223: reserved acts at lines 26–30 and CC decisions at lines 32–36.
• SLSA Build Provenance v1.2, Approved specification, fetched 2026-10-09 UTC: buildDefinition.resolvedDependencies models consumed inputs as ResourceDescriptors with digest fields. This supplies the reference shape; the scoped registration adapter remains a proposal.
• Builder source at ebe6c223: lines 71–105, 183–228, 340–343 and 483–506.
• Source declaration at the same pin.
• Hosted workflows, validation script and readiness tests, inspected 2026-10-09 UTC: no hosted invocation of build_readiness.py --check on committed readiness; the tests at lines 336–354 use the temporary fixture at lines 155–162.
• GitHub Docs: Managing a merge queue and merge_group event, fetched 2026-10-09 UTC: availability, combined candidate and Actions trigger.
• in-toto/attestation Statement v1 at fd2609c16bcb0ac53443e2b4612977f997e8f9a5: subject artifacts and their digests. In the SLSA build model these are outputs.

Evidence-class table

Claim	Evidence class	Command / receipt
Six of ten sampled landings changed the index; main's committed binding was stale after eight	local_integration	Pinned Git-blob measurements ending at ebe6c22; the reproduced-evidence section is byte-identical to f7c68b7
Stale committed readiness has no hosted enforcement	source_review	Workflow and validation-script searches; temporary-root test source; no hosted failure for this stale binding
Consumed-input reference shape is resolvedDependencies; subject identifies the build output	source_review	SLSA v1.2 BuildDefinition, schema and model; pinned in-toto Statement
Queue checks a merge group; current personal-account ownership is outside documented availability	local_integration + source_review	Previously reproduced owner.type=User metadata and dated GitHub documentation retained in the record
A scoped binding could reduce unrelated churn while preserving relevant input integrity	source_review	Proposed direction and inverse; selector completeness and deterministic representation remain untested

Local commands run

timeout 600 nice -n 10 ionice -c2 -n7 python3 scripts/validate.py
rc=0; 70 components, 10,917 hashed files, 239 receipts, 4 profiles

git diff --check
rc=0

timeout 600 nice -n 10 ionice -c2 -n7 bash ~/.local/state/native-agent-stack/coordination/command-center/cc-tools/base_tests_at_head.sh ~/code/native-agent-stack-ns2604-ci-binding-record ebe6c223f65bf20d04c56858e8669e0bd76ef3aa 837cbdd9bb87f77425dfa8b038bdd83c2f20595c
rc=0; no existing test file is modified

Fix logs: coordination/ns2604-coop/lanes/github-ci-finalize-work/ci-binding-record-20261009/{validate-fix,pre-cue-fix}.log on the coordination host. The previous readiness module result at f7c68b7 remains pinned (32 tests, rc 0); it was not rerun for this record-only correction.

Decision record

Readiness evidence binding across landings is a recommendation for CC decision under the cited authority. It recommends less-volatile-binding; its inverse preserves whole-index identity with an accepted finite refresh interval. Replay detecting unrelated churn or missed relevant changes would overturn it. Main was stale after 8 of 10 sampled landings without a hosted failure for the stale binding.

The CC's initial read at f7c68b7 requested one P2 and two P3 corrections; the designated GPT read was ACK with verification gaps. This fix carries all three corrections. Both designated micro reads and required CI must complete on this fix head before the explicit CC landing cue; 5f owns landing. Implementation needs a separate cue. This PR changes one record.

Host evidence

No host receipt files change.

Checklist

• Action pin requirements: no Actions change.
• Workflow permission requirements: no workflows change.
• No secrets are printed, logged or committed; no required secret is added.
• No paid hosting, subscription or billing surface is introduced.
• Peer-owned untracked files and worktrees are preserved.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-09T04:13:36.045204Z	837cbdd	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.