8000 feat(runtime): retain model stop evidence by dominicnunez · Pull Request #216 · dominicnunez/agentos · GitHub
[go: up one dir, main page]

Skip to content

feat(runtime): retain model stop evidence - #216

Merged
dominicnunez merged 6 commits into
mainfrom
feat/model-stop-evidence
Sep 19, 2026
Merged

dominicnunez merged 6 commits into
mainfrom
feat/model-stop-evidence

Conversation

@dominicnunez
Copy link
Copy Markdown
Owner

Problem and behavior

Adaptive planning and model-based intent normalization could remain blocked after shutdown or a security hold without durable stop evidence. This change retains ownership through model return and durable result admission, releases callers after a bounded acknowledgement wait, and records the exact manifest-bound stop request, uncertainty, and eventual local outcome. Returned usage is committed atomically with confirmation; stopped work cannot publish ordinary plans or drafts.

The ledger distinguishes calls that never started from callbacks that actually returned, preserves already committed decisions, and validates retry evidence before a new attempt. Live admission, filtered readers, replay, projection rebuild, and recovery reject contradictory or late evidence and results that omit their manifested execution identity. Genuine unmanifested legacy results remain supported.

Shutdown attempts a bounded drain. A timeout reports failure and retains uncertainty and conservative accounting; local return does not prove remote termination. Suspended-task resumption and the remaining coordination/output/effect and cancellation-timeline audit stay under #178.

Validation

  • Linux build, vet, golangci-lint, dependency consistency, complete repository pre-commit checks, architecture boundaries, and vulnerability scan.
  • Regression and race coverage for blocked calls, late return/usage, storage-write failure, admission boundaries, deadlines, tenant isolation, and real HTTP shutdown/listener failure.
  • Live-versus-replay retry matrix, malformed and contradictory evidence, file reopen/recovery, reserved writers, filtered readers, and storage migration.
  • Representative small/large-history checks; relevant attempt-history validation still grows with that history.

Refs #178

@chatgpt-codex-connector
chatgpt-codex-connector Bot commented Sep 19, 2026 •
Copy link
Copy Markdown

Codex Review Summary

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

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-09-19T23:22:07.749889Z 3fa1c8f New commits
🔒 Security Review ✅ Completed 2026-09-19T23:39:14.687882Z 3fa1c8f Manual request
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.

@dominicnunez
Copy link
Copy Markdown
Owner Author

@codex security review

@chatgpt-codex-connector
Copy link
Copy Markdown

🛡️ Codex Security Review

Security review completed. No security issues were found in this pull request.

Reviewed commit: 3fa1c8f3b5

View security finding report

Only the user who started this review can view the report in Codex.

ℹ️ About Codex security reviews in GitHub

This is an experimental Codex feature. Security reviews are triggered when:

  • You comment "@codex security review"
  • A regular code review gets triggered (for example, "@codex review" or when a PR is opened), and you’re opted in so security review runs alongside code review

Once complete, Codex will leave suggestions, or a comment if no findings are found.

@dominicnunez
dominicnunez merged commit 7248af1 into main Sep 19, 2026
5 checks passed
@dominicnunez
dominicnunez deleted the feat/model-stop-evidence branch September 27, 2026 05:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant

0