Repository navigation
Integrate evidence reliability, atomic correction and offline decision replay - #20
Merged
Merged
Conversation
k2jac9
marked this pull request as ready for review
November 16, 2026 05:49
Codex Review Summary: ✅ CompletedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stored agent results need inspectable eligibility decisions and atomic correction when evidence changes. This integrates the reviewed Rust/SQLite reliability work: a shared current-read explanation, expected-epoch admission batches, a Store-backed semantic fixture, authorized decision export/replay, evaluated policy activation, bounded offline diagnostics, and source-bound runtime qualification.
The original 48 implementation/verification commits are preserved. Review in the seven groups in docs/reliability-review-guide.md; later fixes belong with the implementations they correct.
Hosted verification at
afd8c5d9a65167abdc2a8c8c112416b9a3e91847: all five runtime lanes passed, each with 362 Rust tests, 378 script tests, 18 reference and eight oracle checks, plus explicit integration/example stages. No warnings or test skips appeared in the retained logs. The separate previous-version checkout job passed. PR run · push run.The hosted run exposed ambient library paths in test fixtures, a scheduling-dependent receipt test, and a prepared-worker test that assumed a claim within two seconds. The prepared-worker regression now waits for a real durable claim before advancing only its supervisor test clock. Corrections preserve production refusal and accounting rules; the new refusal regression and deterministic receipt tests passed locally, followed by the full hosted matrix. Checkout pins official v7.0.1 and separately checks v6.1.0; no current upcoming preview was available. Independent review of this integration correction found no actionable issues. See docs/hosted-integration.md for exact boundaries and retained failures.
These are offline contract results, not new model-quality measurements. PR #18's hosted adapter and PR #19's measured report remain separate. This PR performs no live database migration, provider request or model training. The unrecoverable historical superseded preflight display remains documented; newer successful verification does not reconstruct it.