What's Changed

• build(deps): bump the github-actions group with 4 updates by @dependabot[bot] in #73
• build(deps): bump @openai/codex from 0.160.0 to 0.162.1 in the npm-dependencies group by @dependabot[bot] in #74
• feat(ai-review): suppress repeated findings after a maintainer verdict by @hln33 in #75

Full Changelog: v0.4.0...v0.5.0

Highlights

• Configure the runner for every reusable workflow with the optional runs-on input; the default remains ubuntu-latest.
• Use GPT-6 Astra for Codex workflows and GPT-6 Luna for notification summaries, with configurable model and region inputs.
• Return answered needs-info issues to triage and preserve successful Claude reviews while retaining the CLI's 20-round limit.
• Update Codex CLI to 0.160.0 and Claude Code Action to 1.0.242.

Upgrade notes

The default Codex model is now openai.gpt-6-astra in us-west-2; notification summaries use openai.gpt-6-luna in us-east-1. Ensure the configured Bedrock role can invoke the selected model in its region, or explicitly set the workflow's model and region inputs.

Custom runners used for model jobs must be ephemeral and provide passwordless sudo, adduser, and setfacl. See the runner requirements.

What's Changed

• fix(stale-issue-closer): add needs-triage after clearing needs-info by @hln33 in #53
• build(deps): bump anthropics/claude-code-action from 1.0.211 to 1.0.217 in the github-actions group by @dependabot[bot] in #60
• build(deps): bump @openai/codex from 0.151.0 to 0.153.4 in the npm-dependencies group by @dependabot[bot] in #59
• build(deps): bump the github-actions group with 2 updates by @dependabot[bot] in #62
• build(deps): bump @openai/codex from 0.153.4 to 0.155.1 in the npm-dependencies group by @dependabot[bot] in #61
• build(deps): bump @openai/codex from 0.155.1 to 0.158.0 in the npm-dependencies group by @dependabot[bot] in #65
• build(deps): bump anthropics/claude-code-action from 1.0.231 to 1.0.236 in the github-actions group by @dependabot[bot] in #66
• feat: add runs-on input to the shareable workflows by @wangyb-A in #67
• build(deps): bump @openai/codex from 0.158.0 to 0.160.0 in the npm-dependencies group by @dependabot[bot] in #69
• build(deps): bump anthropics/claude-code-action from 1.0.236 to 1.0.242 in the github-actions group by @dependabot[bot] in #68
• Update Codex workflows to GPT-6 models by @zhongkechen in #63
• fix: enforce Claude review turn limit in CLI wrapper by @zhongkechen in #72

New Contributors

• @wangyb-A made their first contribution in #67

Full Changelog: v0.3.1...v0.4.0

What's Changed

• docs: add AI-assisted developer workflow guide by @zhongkechen in #52
• build(deps): bump anthropics/claude-code-action from 1.0.199 to 1.0.202 in the github-actions group by @dependabot[bot] in #55
• build(deps): bump the github-actions group with 2 updates by @dependabot[bot] in #57
• build(deps): bump @openai/codex from 0.149.1 to 0.151.0 in the npm-dependencies group by @dependabot[bot] in #56
• Disable Claude reviews by default by @zhongkechen in #58

Full Changelog: v0.3.0...v0.3.1

Highlights

Durable AI review telemetry

• Adds stable review, finding, observation, and suggestion identifiers so findings can be correlated across reruns and pull request revisions.
• Records exact event timestamps and reviewed base/head revisions on the orphan ai-review-telemetry-v1 branch.
• Supports structured maintainer verdicts: accepted, deferred, false-positive, out-of-scope, and already-fixed.
• Tracks human review activity and exact suggestion adoption, with canonical NDJSON export for analysis.

Expanded AI commands and workflows

• Adds the /ai implement issue workflow and team-triggered /ai review commands.
• Allows maintainers to provide per-review guidance.
• Separates Codex issue implementation from PR review and simplifies PR review continuation setup.
• Improves generated Codex PR descriptions and safely supports authorized workflow changes.

Customization

• Supports configurable runtime environments for consuming repositories.

Consumer upgrade note

Consumers updating their pinned AI review workflow to v0.3.0 must change the reusable workflow job permission from contents: read to contents: write. The trusted telemetry jobs use that permission to maintain the orphan telemetry branch; model jobs remain explicitly restricted to contents: read.

To capture human-review activity and support inline verdict replies, callers should also subscribe to the pull_request_review and pull_request_review_comment events documented in the AI review setup guide.

What Changed

• Remove Claude review retry and reduce timeout (#22)
• Default Claude reviews to Sonnet 5 (#23)
• Add stale issue closer workflow (#26, #28)
• Add AI-command issue implementation workflow (#32)
• Support team-triggered AI PR reviews (#33)
• Fix Codex implementation workspace access (#34)
• Update Codex CLI and sandbox preflight (#35)
• Enable Dependabot dependency updates (#36, #37)
• Normalize Codex credential expiration (#38)
• Allow workflow changes from /ai implement (#39)
• Default Codex issue implementation reasoning to xhigh (#40)
• Allow two-hour Codex runs with credential refresh (#41)
• Allow new implementation PRs after base advances (#42)
• Allow Codex to push authorized workflow changes (#43)
• Allow consumers to override the runtime environment name (#44)
• Allow per-review guidance in /ai review commands (#45)
• Improve generated Codex PR descriptions (#47)
• Split Codex issue and PR review workflows (#48)
• Rename agent-neutral workflows for AI (#49)
• Simplify AI PR review address caller setup (#50)
• Add durable AI review telemetry (#51)

New Contributors

• @hln33 made their first contribution in #26.
• @dependabot[bot] made its first contribution in #37.

Full changelog: v0.2.0...v0.3.0

What's Changed

• docs: clarify AI review runtime approval setup by @zhongkechen in #12
• Add hardened AI issue triage by @zhongkechen in #13
• Add AI summaries to Slack notifications by @zhongkechen in #14
• Align issue triage with repository label taxonomy by @zhongkechen in #19
• Check OpenTelemetry spec compliance in AI reviews by @zhongkechen in #20
• Use GPT-5.6 Luna for Slack notification summaries by @zhongkechen in #18

Full Changelog: v0.1.0...v0.2.0

What's Changed

• Add reusable AI PR review workflow by @zhongkechen in #1
• Add reusable Slack notification workflow by @zhongkechen in #2
• Document AI review environment secret inheritance by @zhongkechen in #3
• Harden AI review workflows and publish validated suggestions by @zhongkechen in #5
• Add workflow to auto-label new issues with needs-triage by @ayushiahjolia in #6
• Refactor AI review workflows consistently by @zhongkechen in #8
• fix(ci): Prevent script injection in ai-pr-review workflow by @ayushiahjolia in #10
• Consolidate AI review workflows and documentation by @zhongkechen in #9

New Contributors

• @zhongkechen made their first contribution in #1
• @ayushiahjolia made their first contribution in #6

Full Changelog: https://github.com/aws/aws-durable-execution-ci/commits/v0.1.0