Remove dedicated fixers, bump to v0.58.1#687
Conversation
github-actions
bot
added
the
big_boom
- Fix remaining Code Simplifier/Bug Exterminator/Test Improver references in blog posts (welcome, quality crew, day in factory, silence) - Add severity-threshold and title-prefix to code-complexity-detector README - Fix safe outputs description (expires vs auto-closes) - Fix broken link in scheduled-fix README - Soften "each detector" claim in chaining docs - Use curl -fsSL for agentics-maintenance install Made-with: Cursor
Resolve conflicts with #659 (assign-to-Copilot chaining option). Keep our fixer-removal changes; incorporate the new "assign to Copilot" section and Copilot row in the chaining comparison table. Made-with: Cursor
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (80)
📝 WalkthroughWalkthroughThe PR introduces a computed per-run artifact prefix (artifact_prefix) propagated from activation to downstream jobs and used to dynamically name artifacts across many workflows. It standardizes safe-outputs handling (replacing some uploads with copy steps and new safe-outputs tooling/configs), adds phase/version environment signals (GH_AW_PHASE, GH_AW_VERSION, GITHUB_AW), updates many action and image pins (notably AWF 0.23.0 → 0.24.1 and setup action v0.57.0 → v0.58.1), adds cross-repo setup guidance, and removes or renames several workflow docs/workflows. Possibly related PRs
✨ Finishing Touches🧪 Generate unit tests (beta)
📝 Coding Plan
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 6
Note
Due to the large number of review comments, Critical severity comments were prioritized as inline comments.
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (4)
.github/workflows/gh-aw-framework-best-practices.lock.yml (1)
149-159:⚠️ Potential issue | 🟠 MajorUse a real repo source before emitting the cross-repo token hint.
Line 155 references
steps.resolve-host-repo.outputs.target_repo, butactivationhas noresolve-host-repostep. With the currentfailure()guard, this hint can fire after unrelated activation failures and send maintainers toward the wrong remediation. Gate it onsteps.validate-secret.outcome == 'failure'and compare against an actually resolved caller/host repo.Based on learnings,
.github/workflows/*.lock.ymlfiles are generated bygh aw compile, so this fix should be made in the corresponding source.mdworkflow.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-framework-best-practices.lock.yml around lines 149 - 159, The cross-repo hint currently checks failure() and references steps.resolve-host-repo.outputs.target_repo which may not exist; change the conditional to explicitly gate the "Cross-repo setup guidance" job on the validate-secret step failing (e.g., if: steps.validate-secret.outcome == 'failure') and then only compare against a resolved caller/host repo output that is actually defined (ensure steps.resolve-host-repo exists in the source workflow or use the correct step id that produces a target_repo output); update the source .md workflow that generates this .lock.yml (not the compiled file) so the compiled action includes the corrected if condition and valid step reference..github/workflows/agent-deep-dive.lock.yml (1)
649-700:⚠️ Potential issue | 🟡 MinorRemove the
create_discussionbranch from this tool schema.This workflow only exposes
create_issue,missing_tool,missing_data, andnoop, and the reporting prompt later requires filing an issue. Telling the model to usecreate_discussioncreates an impossible path that can turn a real finding into amissing_tool/noopinstead of the expected issue.Based on learnings,
.github/workflows/*.lock.ymlfiles are generated bygh aw compile, so this should be fixed in the source.mdand recompiled.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/agent-deep-dive.lock.yml around lines 649 - 700, The tool schema currently includes a branch for "create_discussion" which must be removed so the workflow only exposes "create_issue", "missing_tool", "missing_data", and "noop"; remove the "create_discussion" entry from the tools array/schema (where "name": "create_discussion" appears) and ensure any descriptions or prompts that reference "create_discussion" are updated to require "create_issue" instead, then update the original source markdown that generates this lockfile (the .md that defines the tools) and recompile the lockfile with gh aw compile so the generated .github/workflows/*.lock.yml no longer contains the create_discussion branch..github/workflows/agent-efficiency.lock.yml (1)
1056-1075:⚠️ Potential issue | 🟠 MajorRedact the new artifact files before uploading them.
safeoutputs.jsonlandagent_output.jsonare created after Line 1038'sRedact secrets in logsstep, then uploaded in Lines 1125-1126. Any secret that reaches those files now bypasses the job's only explicit scrub pass. Move the copy/ingest steps before redaction, or run a second redaction pass immediately beforeupload-artifact.Also applies to: 1114-1126
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/agent-efficiency.lock.yml around lines 1056 - 1075, The uploaded artifacts (safeoutputs.jsonl and agent_output.json) are created after the "Redact secrets in logs" step, so either move the "Copy Safe Outputs" and "Ingest agent output" steps to run before the "Redact secrets in logs" step or add a second redaction pass immediately before the existing upload-artifact steps (the steps that upload safeoutputs.jsonl/agent_output.json); update the "Copy Safe Outputs" and "Ingest agent output" steps (and the collect_output action invocation) to ensure they operate on the redacted files or invoke the same redaction logic used by "Redact secrets in logs" so no secrets reach the upload-artifact steps..github/workflows/gh-aw-estc-pr-buildkite-detective.lock.yml (1)
1357-1371:⚠️ Potential issue | 🟠 MajorThe
GH_AW_FAILURE_REPORT_AS_ISSUEflag should not appear only in the.lock.ymlfile.The flag is present in
.github/workflows/gh-aw-estc-pr-buildkite-detective.lock.ymlbut absent from the source.github/workflows/gh-aw-estc-pr-buildkite-detective.mdfile. Since.lock.ymlfiles are compiled outputs of.mdsources viagh aw compile, this configuration must be defined in the.mdsource file for reproducibility and maintainability.Additionally, the source
.mddeclarespermissions: {actions: read, contents: read, issues: read, pull-requests: read}(all read-only), which lacksissues: write. If the flag triggers issue creation in the externalhandle_agent_failure.cjsscript, the workflow will fail with insufficient permissions. Update the.mdsource file to include both the flag definition and the requiredissues: writepermission.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-estc-pr-buildkite-detective.lock.yml around lines 1357 - 1371, Add the GH_AW_FAILURE_REPORT_AS_ISSUE environment variable to the source workflow file (gh-aw-estc-pr-buildkite-detective.md) so it is compiled into the .lock.yml (match the name GH_AW_FAILURE_REPORT_AS_ISSUE used in the handle_agent_failure step), and update the workflow permissions block in that same .md file to include issues: write (in addition to the existing read permissions) so the external handle_agent_failure.cjs can create issues when the flag is enabled.
🟠 Major comments (16)
.github/workflows/gh-aw-create-pr-from-issue.lock.yml-1346-1352 (1)
1346-1352:⚠️ Potential issue | 🟠 MajorUse the prefixed artifact name in the conclusion job.
Line 1351 still downloads
agent, but the agent job now uploads${{ needs.activation.outputs.artifact_prefix }}agent. When the prefix is non-empty,agent_output.jsonnever reaches the conclusion job, so noop/failure handling loses the agent state. The same stale literal appears in the other updated workflows in this PR.Based on learnings, `.github/workflows/*.lock.yml` files are generated by `gh aw compile`, so fix the source workflow/fragment and recompile.Suggested fix
- name: agent + name: ${{ needs.activation.outputs.artifact_prefix }}agent🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-create-pr-from-issue.lock.yml around lines 1346 - 1352, The conclusion job currently downloads a literal artifact name "agent" (job id download-agent-output) but the activation job uploads "${{ needs.activation.outputs.artifact_prefix }}agent", so when a prefix is set the artifact is never found; update the artifact name in the download step to use the same prefixed variable (e.g. set the download step's name field to use ${{ needs.activation.outputs.artifact_prefix }}agent) in the source workflow/fragment that generates these .lock.yml files and then re-run gh aw compile to regenerate the .github/workflows/*.lock.yml outputs so all workflows (including the other updated workflows in this PR) use the prefixed artifact name consistently..github/workflows/gh-aw-code-complexity-detector.lock.yml-1503-1515 (1)
1503-1515:⚠️ Potential issue | 🟠 MajorDownload the prefixed agent artifact here too.
Line 1344 uploads
${{ needs.activation.outputs.artifact_prefix }}agent, and Line 1644 already downloads that prefixed name insafe_outputs, but this job still requests plainagenton Line 1508. Whenartifact_prefixis non-empty,GH_AW_AGENT_OUTPUTnever gets populated here, so noop/failure handling loses the persisted agent output.Based on learnings, persist outcomes through safe outputs because uncaptured local state is lost after the run, and these
.lock.ymlfiles are generated from the corresponding.mdsources.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-code-complexity-detector.lock.yml around lines 1503 - 1515, The download step with id download-agent-output is fetching the hardcoded artifact name "agent" which mismatches the uploaded artifact name that uses needs.activation.outputs.artifact_prefix; update the download action to request the prefixed artifact name (concatenate needs.activation.outputs.artifact_prefix with "agent") so the artifact is found whenever a prefix is present, and ensure the subsequent Setup agent output environment variable step still writes GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json into $GITHUB_ENV (and preserves the same /tmp/gh-aw/ path) so the job's safe_outputs flow receives the persisted agent output..github/workflows/gh-aw-estc-pr-buildkite-detective.lock.yml-1316-1328 (1)
1316-1328:⚠️ Potential issue | 🟠 MajorDownload the prefixed agent artifact in
conclusion.The agent artifact is now uploaded as
${{ needs.activation.outputs.artifact_prefix }}agent, but this step still downloadsagent. That breaks the handoff ofagent_output.jsoninto the conclusion phase, so noop/failure handling runs without the agent output it expects.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-estc-pr-buildkite-detective.lock.yml around lines 1316 - 1328, The Download agent output artifact step (id: download-agent-output) currently requests the static artifact name "agent" but the uploader uses the dynamic prefix from needs.activation.outputs.artifact_prefix; change the artifact name in that step to "${{ needs.activation.outputs.artifact_prefix }}agent" so the correct prefixed artifact is downloaded, keeping the rest of the step (path /tmp/gh-aw/ and the follow-up step that sets GH_AW_AGENT_OUTPUT) unchanged..github/workflows/gh-aw-bug-hunter.lock.yml-1382-1394 (1)
1382-1394:⚠️ Potential issue | 🟠 MajorDownload the prefixed agent artifact in
conclusion.Earlier in this workflow the agent job uploads
${{ needs.activation.outputs.artifact_prefix }}agent, andsafe_outputsalready downloads that name. Keepingconclusiononagentmeansagent_output.jsonnever lands under/tmp/gh-aw, so noop/missing-tool/failure handling loses the agent output for this run.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-bug-hunter.lock.yml around lines 1382 - 1394, The workflow currently downloads the hardcoded artifact name "agent" in the "download-agent-output" step causing the prefixed upload (from needs.activation.outputs.artifact_prefix) to be missed; update the download step's artifact name to use the uploaded prefix (i.e. "${{ needs.activation.outputs.artifact_prefix }}agent") so the file lands in /tmp/gh-aw/ and GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json is correctly set; keep the step id "download-agent-output" and the subsequent env export to ensure downstream checks (noop/missing-tool/failure handling) read the prefixed agent_output.json..github/workflows/gh-aw-information-architecture.lock.yml-1380-1392 (1)
1380-1392:⚠️ Potential issue | 🟠 MajorDownload the prefixed agent artifact in
conclusion.The agent artifact is now uploaded as
${{ needs.activation.outputs.artifact_prefix }}agent, but this step still requestsagent. That leavesGH_AW_AGENT_OUTPUTunset here, so noop/failure processing no longer sees the agent output for the current run.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-information-architecture.lock.yml around lines 1380 - 1392, The download step (id: download-agent-output) currently requests the literal artifact name "agent" but the artifact is uploaded as "${{ needs.activation.outputs.artifact_prefix }}agent", so GH_AW_AGENT_OUTPUT never gets set; update the download-agent-output step to use the prefixed artifact name from needs.activation.outputs.artifact_prefix (so it downloads "${{ needs.activation.outputs.artifact_prefix }}agent") and keep the subsequent Setup agent output environment variable step that writes GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json to GITHUB_ENV unchanged so the environment variable is reliably populated..github/workflows/gh-aw-breaking-change-detector.lock.yml-1387-1399 (1)
1387-1399:⚠️ Potential issue | 🟠 MajorConclusion still downloads the pre-prefix artifact name.
Line 1392 requests
agent, but Line 1228 uploads${{ needs.activation.outputs.artifact_prefix }}agent, and Line 1528 already consumes the prefixed name insafe_outputs. Because this step iscontinue-on-error: true, the miss stays silent andGH_AW_AGENT_OUTPUTnever gets set on Line 1399, sonoop,missing_tool, and failure handling lose the agent output. I see the same stale literal in the other prefixed conclusion jobs in this review set, so this looks like a shared source/template miss. Please fix it in the source workflow/fragment and recompile. Based on learnings, files under.github/workflowswith the.lock.ymlextension are auto-generated by runninggh aw compilefrom corresponding.mdsource files.Suggested fix
- name: agent + name: ${{ needs.activation.outputs.artifact_prefix }}agent🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-breaking-change-detector.lock.yml around lines 1387 - 1399, The conclusion job is downloading the hardcoded artifact name "agent" (step id download-agent-output) instead of the prefixed name used when uploading (${ { needs.activation.outputs.artifact_prefix } }agent), so replace the literal artifact name with the same prefixed expression and ensure the conditional that sets GH_AW_AGENT_OUTPUT uses the successful download step (download-agent-output) as before; update the template/source fragment that generates this workflow (the file that sets artifact_prefix and safe_outputs) and re-run the generator (gh aw compile) so the compiled .lock.yml emits the prefixed artifact name and GH_AW_AGENT_OUTPUT is correctly set when the download succeeds..github/workflows/gh-aw-dependency-review.lock.yml-150-155 (1)
150-155:⚠️ Potential issue | 🟠 MajorDon't gate this hint on an undefined step output.
activationnever definesresolve-host-repo, so thisif:has no real cross-repo signal and can emit the caller-secret guidance for unrelated activation failures. Restore the missing resolver or switch this check to a context that actually exists.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-dependency-review.lock.yml around lines 150 - 155, The "Cross-repo setup guidance" step is gated on an undefined step output (resolve-host-repo), so the hint can appear for unrelated failures; either add/restore the missing resolver step named resolve-host-repo that emits outputs.target_repo before this step, or change the step's if condition to use an existing context/step (for example an existing step output or github.repository comparison) that actually indicates a cross-repo invocation; update the "Cross-repo setup guidance" step's if to reference the restored resolve-host-repo outputs.target_repo or a valid existing context so the guidance only runs for true cross-repo cases..github/workflows/gh-aw-code-quality-audit.lock.yml-153-158 (1)
153-158:⚠️ Potential issue | 🟠 MajorDon't gate this hint on an undefined step output.
activationnever definesresolve-host-repo, so thisif:has no real cross-repo signal and can emit the caller-secret guidance for unrelated activation failures. Restore the missing resolver or switch this check to a context that actually exists.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-code-quality-audit.lock.yml around lines 153 - 158, The conditional that gates the "Cross-repo setup guidance" uses a non-existent step output steps.resolve-host-repo.outputs.target_repo, so the hint can appear for unrelated failures; either restore the resolver step that produces resolve-host-repo.outputs.target_repo (e.g., re-add the resolve-host-repo step that sets target_repo) or change the if: expression to a valid context such as comparing github.repository to the caller repository (for example use github.repository != github.event.repository or another existing step/output) so the hint only runs for true cross-repo invocations..github/workflows/gh-aw-internal-gemini-cli.lock.yml-1367-1376 (1)
1367-1376:⚠️ Potential issue | 🟠 MajorMirror the prefixed artifact name in conclusion.
Line 1205 uploads
${{ needs.activation.outputs.artifact_prefix }}agent, but Line 1369 still requestsagent. That mismatch means conclusion no longer tracks the artifact produced earlier in the workflow, and a miss here leavesGH_AW_AGENT_OUTPUTunset at Line 1376.Suggested fix
- name: agent + name: ${{ needs.activation.outputs.artifact_prefix }}agent🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-internal-gemini-cli.lock.yml around lines 1367 - 1376, The workflow downloads and sets GH_AW_AGENT_OUTPUT using the hard-coded artifact name "agent", but the upload used the prefixed name from needs.activation.outputs.artifact_prefix; update the download-and-env step (where steps.download-agent-output is referenced and GH_AW_AGENT_OUTPUT is set) to request the same prefixed artifact by replacing the artifact name "agent" with "${{ needs.activation.outputs.artifact_prefix }}agent" so the downloaded file path and the GH_AW_AGENT_OUTPUT value point to the actual uploaded artifact..github/workflows/gh-aw-code-quality-audit.lock.yml-1407-1416 (1)
1407-1416:⚠️ Potential issue | 🟠 MajorMirror the prefixed artifact name in conclusion.
Line 1245 uploads
${{ needs.activation.outputs.artifact_prefix }}agent, but Line 1409 still requestsagent. That mismatch means this job no longer follows the new artifact naming contract, and a miss here leavesGH_AW_AGENT_OUTPUTunset at Line 1416.Suggested fix
- name: agent + name: ${{ needs.activation.outputs.artifact_prefix }}agent🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-code-quality-audit.lock.yml around lines 1407 - 1416, The download step and GH_AW_AGENT_OUTPUT should use the same prefixed artifact name as the upload; update the download-artifact `name` value from the literal "agent" to use the activation output prefix (i.e. "${{ needs.activation.outputs.artifact_prefix }}agent") and ensure the path and subsequent GH_AW_AGENT_OUTPUT value reference the matching file (e.g. /tmp/gh-aw/${{ needs.activation.outputs.artifact_prefix }}agent_output.json) so the find and echo lines point to the correct prefixed artifact file..github/workflows/gh-aw-dependency-review.lock.yml-1434-1443 (1)
1434-1443:⚠️ Potential issue | 🟠 MajorMirror the prefixed artifact name in conclusion.
Line 1271 uploads
${{ needs.activation.outputs.artifact_prefix }}agent, but Line 1436 still requestsagent. That mismatch means this step no longer follows the upload contract, and a miss here leavesGH_AW_AGENT_OUTPUTunset at Line 1443.Suggested fix
- name: agent + name: ${{ needs.activation.outputs.artifact_prefix }}agent🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-dependency-review.lock.yml around lines 1434 - 1443, The download step and the subsequent "Setup agent output environment variable" assume the artifact name is literally "agent", but the upload used "${{ needs.activation.outputs.artifact_prefix }}agent"; update the download step's artifact name (used by actions/download-artifact and referenced by steps.download-agent-output) to mirror the upload by using "${{ needs.activation.outputs.artifact_prefix }}agent", and ensure the GH_AW_AGENT_OUTPUT environment variable (written to $GITHUB_ENV in the "Setup agent output environment variable" step) still points to /tmp/gh-aw/agent_output.json so the rest of the workflow reads the correct file when the prefixed artifact is downloaded..github/workflows/gh-aw-internal-gemini-cli.lock.yml-159-164 (1)
159-164:⚠️ Potential issue | 🟠 MajorThis failure hint points at the wrong thing.
activationnever definesresolve-host-repo, so this branch has no real cross-repo signal, and the message tells callers to setCOPILOT_GITHUB_TOKENeven though Line 154 validatesGEMINI_API_KEY.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-internal-gemini-cli.lock.yml around lines 159 - 164, The failure hint references a nonexistent step id resolve-host-repo and tells callers to set COPILOT_GITHUB_TOKEN even though activation never defines that step and the workflow validates GEMINI_API_KEY at Line 154; update this branch to either use the correct cross-repo signal step id (or add a resolve-host-repo step) or change the error text to reference GEMINI_API_KEY and the actual condition used by activation so the hint matches real validation logic (adjust the condition, step id, or message in the Cross-repo setup guidance block accordingly)..github/workflows/gh-aw-deep-research.lock.yml-1345-1348 (1)
1345-1348:⚠️ Potential issue | 🟠 MajorUse the prefixed agent artifact in
conclusion.Line 1347 still requests
agent, but the agent job uploads${{ needs.activation.outputs.artifact_prefix }}agent. That leavesGH_AW_AGENT_OUTPUTunset for the noop and failure handlers.Based on learnings, outcomes must be persisted through safe outputs in GitHub Actions runs because local state is lost, and
.lock.ymlfiles here are generated bygh aw compile, so fix this in the source.mdor shared fragment.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-deep-research.lock.yml around lines 1345 - 1348, The workflow downloads the artifact named "agent" but the activation job uploads it as "${{ needs.activation.outputs.artifact_prefix }}agent", which leaves GH_AW_AGENT_OUTPUT unset for noop/failure handlers; update the download-artifact step to use the prefixed artifact name matching the upload (i.e., replace the static "agent" name with the same expression used by the uploader: ${ { needs.activation.outputs.artifact_prefix } }agent) so the artifact path and GH_AW_AGENT_OUTPUT are correctly set for all handlers and persisted via the safe outputs mechanism..github/workflows/gh-aw-estc-docs-pr-review.lock.yml-1479-1482 (1)
1479-1482:⚠️ Potential issue | 🟠 MajorUse the prefixed agent artifact in
conclusion.Line 1481 still requests
agent, but the agent job now uploads${{ needs.activation.outputs.artifact_prefix }}agent. If this misses, the later noop and failure handlers never receiveagent_output.json.Based on learnings, outcomes must be persisted through safe outputs in GitHub Actions runs because local state is lost, and
.lock.ymlfiles here are generated bygh aw compile, so fix this in the source.mdor shared fragment.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-estc-docs-pr-review.lock.yml around lines 1479 - 1482, The workflow step currently downloads the unprefixed artifact name "agent" but the activation job uploads a prefixed artifact using needs.activation.outputs.artifact_prefix, so change the download artifact name to use the same prefix (i.e. name: ${{ needs.activation.outputs.artifact_prefix }}agent) and ensure subsequent steps that expect agent_output.json still read from /tmp/gh-aw/; also update the source template/markdown or shared fragment that generates this lockfile so the prefixed artifact name is used everywhere (preserving the artifact_prefix reference) so noop and failure handlers receive the correct artifact..github/workflows/gh-aw-duplicate-issue-detector.lock.yml-1266-1269 (1)
1266-1269:⚠️ Potential issue | 🟠 MajorUse the prefixed agent artifact in
conclusion.Line 1268 still requests
agent, but the agent job now uploads${{ needs.activation.outputs.artifact_prefix }}agent. If this misses, the "no duplicate found" noop path and failure handling loseagent_output.json.Based on learnings, outcomes must be persisted through safe outputs in GitHub Actions runs because local state is lost, and
.lock.ymlfiles here are generated bygh aw compile, so fix this in the source.mdor shared fragment.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-duplicate-issue-detector.lock.yml around lines 1266 - 1269, Update the artifact download step that currently uses name: agent to use the prefixed artifact name from the activation output (e.g. name: ${{ needs.activation.outputs.artifact_prefix }}agent) so the "conclusion" path receives the correct artifact (agent_output.json); locate the actions/download-artifact usage block and change the name value accordingly in the source fragment or markdown that generates this .lock.yml so all branches (noop and failure) persist the prefixed agent artifact via outputs..github/workflows/gh-aw-deep-research.lock.yml-162-167 (1)
162-167:⚠️ Potential issue | 🟠 MajorFix the Gemini cross-repo remediation path.
Line 163 references
steps.resolve-host-repo.outputs.target_repo, but this job has noresolve-host-repostep. Lines 165-166 tell callers to configureCOPILOT_GITHUB_TOKENeven though this workflow validatesGEMINI_API_KEY, so cross-repo secret failures will point users to the wrong fix.This
.lock.ymlis auto-generated bygh aw compilefrom its source.mdfile; fix in the source file or shared fragment instead.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-deep-research.lock.yml around lines 162 - 167, The failure message block incorrectly references a non-existent step name resolve-host-repo and directs callers to set COPILOT_GITHUB_TOKEN even though this workflow validates GEMINI_API_KEY; update the source fragment or .md that generates this lock file so the cross-repo remediation uses the correct outputs and secret name: remove or replace references to steps.resolve-host-repo.outputs.target_repo, and change the error guidance to mention GEMINI_API_KEY (and the correct secret configuration steps for workflow_call cross-repo), ensuring the generated .lock.yml contains the corrected step name/condition and error messages.
🟡 Minor comments (4)
.github/workflows/agent-efficiency.lock.yml-1275-1287 (1)
1275-1287:⚠️ Potential issue | 🟡 MinorGuard
GH_AW_AGENT_OUTPUTin both consumers.A successful artifact download here only proves the
agentartifact exists. It does not guaranteeagent_output.jsonexists inside it, so exporting the path unconditionally can turn a missing output into a secondary handler failure.Suggested guard for both blocks
- echo "GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json" >> "$GITHUB_ENV" + if [ -f /tmp/gh-aw/agent_output.json ]; then + echo "GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json" >> "$GITHUB_ENV" + fiAlso applies to: 1383-1395
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/agent-efficiency.lock.yml around lines 1275 - 1287, The workflow currently sets GH_AW_AGENT_OUTPUT unconditionally after the download-agent-output step, which only proves the artifact exists, not that agent_output.json exists inside it; update the download-agent-output block (step id download-agent-output) to check that /tmp/gh-aw/agent_output.json actually exists before echoing GH_AW_AGENT_OUTPUT to $GITHUB_ENV, and update any consumer steps that rely on GH_AW_AGENT_OUTPUT to first guard on the environment variable being set and that the file exists (e.g., replace bare references to GH_AW_AGENT_OUTPUT with a conditional if: env.GH_AW_AGENT_OUTPUT != '' && <file-exists-check> or an equivalent run-time check) so consumers only proceed when the file is present..github/workflows/gh-aw-create-pr-from-issue.lock.yml-160-165 (1)
160-165:⚠️ Potential issue | 🟡 MinorFix dangling step reference in the source workflow.
Line 161 references
steps.resolve-host-repo.outputs.target_repo, but the "Cross-repo setup guidance" step lacks the requiredid: resolve-host-repo. This causes the condition to fail at runtime when checking repository context. Locate the source.mdfile or compilation template that generates this step, add the missingid, and recompile withgh aw compile.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-create-pr-from-issue.lock.yml around lines 160 - 165, The "Cross-repo setup guidance" step references steps.resolve-host-repo.outputs.target_repo but the step generating that output is missing the id; locate the source template or .md that emits the step which should produce outputs.target_repo (the step referenced as resolve-host-repo), add the missing id: resolve-host-repo to that step so the outputs are addressable, then recompile the workflow with gh aw compile to regenerate the YAML; ensure the id name exactly matches the referenced symbol steps.resolve-host-repo so the if condition works at runtime..github/workflows/gh-aw-code-complexity-detector.lock.yml-164-169 (1)
164-169:⚠️ Potential issue | 🟡 MinorUse a defined step output in the cross-repo hint.
Line 165 checks
steps.resolve-host-repo.outputs.target_repo, but this job never defines a step withid: resolve-host-repo. The new guidance step therefore cannot reliably distinguish cross-repo failures from local ones. Restore the missing resolver step or switch this condition to context that already exists in the workflow.Based on learnings, files under
.github/workflows/*.lock.ymlare generated bygh aw compile, so this should be fixed in the source.mdworkflow or shared fragment.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-code-complexity-detector.lock.yml around lines 164 - 169, The conditional uses steps.resolve-host-repo.outputs.target_repo but no step with id "resolve-host-repo" exists, so update the workflow source to either add/restore a step with id: resolve-host-repo that sets outputs.target_repo (the resolver that detects the calling repo) or change the if condition to rely on an existing context/output (for example compare github.repository with github.event.repository.full_name or another defined step output) so the Cross-repo setup guidance step can reliably detect cross-repo calls; adjust the source .md workflow/fragment used by gh aw compile accordingly..github/workflows/gh-aw-estc-docs-patrol-external.lock.yml-153-158 (1)
153-158:⚠️ Potential issue | 🟡 Minor
resolve-host-repostep is referenced but never defined.Line 154 references
steps.resolve-host-repo.outputs.target_repo, but no step with that id exists in this job. The cross-repo guidance conditional will fail. Restore the missing step in the source workflow or replace the guard with an available output/context.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-estc-docs-patrol-external.lock.yml around lines 153 - 158, The conditional references a non-existent step id steps.resolve-host-repo.outputs.target_repo (resolve-host-repo) so the job will fail; either restore the missing step that sets outputs.target_repo (e.g., add back the resolve-host-repo step that determines target_repo) or change the conditional to use an existing output/context (for example github.repository, github.event.repository.full_name, or the actual step id that resolves the host repo) so the if: expression points to a valid output; update the job to ensure the step id and output name exactly match the conditional (resolve-host-repo and outputs.target_repo) if you choose to restore it.
🧹 Nitpick comments (1)
.github/workflows/gh-aw-code-complexity-detector.md (1)
143-146: File path handling could break with spaces.The
xargs wc -lcommand may fail on paths containing spaces or special characters.🛡️ More robust alternatives
Option 1: Use null-delimited input:
printf '%s\0' $FILES | xargs -0 wc -l | sort -nr | head -200 | awk 'NR>1 {print $2}'Option 2: Avoid xargs entirely:
echo "$FILES" | tr ' ' '\n' | while IFS= read -r f; do wc -l "$f" 2>/dev/null; done | sort -nr | head -200 | awk '{print $2}'🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/gh-aw-code-complexity-detector.md around lines 143 - 146, Replace the fragile pipeline "printf '%s\n' "$FILES" | xargs wc -l | sort -nr | head -200 | awk '{print $2}'" with a null-delimited safe variant: produce null-separated filenames from the $FILES variable and use xargs -0 when invoking wc -l (and adjust the awk expression to ignore the total line), or alternatively iterate over each filename with a while IFS= read -r loop and call wc -l "$f" to avoid xargs entirely; update the pipeline accordingly wherever the original command appears so paths with spaces or special characters are handled safely.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In @.github/workflows/gh-aw-agent-suggestions.lock.yml:
- Around line 1375-1381: The download step "download-agent-output" uses a
hardcoded artifact name "agent" but the upload used the prefixed name from
needs.activation.outputs.artifact_prefix (the upload at the "conclusion" job
used `${{ needs.activation.outputs.artifact_prefix }}agent`), causing mismatched
artifact names; update the download step's with.name to use the same prefix
expression (combine needs.activation.outputs.artifact_prefix with "agent") so
the actions/download-artifact call requests the identical artifact name produced
by the upload.
In @.github/workflows/gh-aw-docs-drift.lock.yml:
- Around line 1403-1406: The download step uses a hardcoded artifact name
"agent" but the upload uses the prefixed name `${{
needs.activation.outputs.artifact_prefix }}agent`, so update the source .md that
generates this workflow to use the same prefixed name when downloading (i.e.,
reference `${{ needs.activation.outputs.artifact_prefix }}agent` in the
download-artifact `name` field instead of `agent`) so the generated .lock.yml
contains the matching artifact name; regenerate the workflow after making the
change.
- Around line 159-164: The condition uses
steps.resolve-host-repo.outputs.target_repo but there is no step with id
"resolve-host-repo"; either add or rename a step to expose outputs.target_repo
(e.g., create a step with id: resolve-host-repo that sets output "target_repo")
or change the if condition to reference an existing step/output (for example
replace steps.resolve-host-repo.outputs.target_repo with the correct output name
from the step that resolves the host repo or compare directly to
github.repository). Update the workflow so the if: expression points to a valid
step id/output (resolve-host-repo or the actual existing step id) to ensure the
cross-repo guidance runs correctly.
In @.github/workflows/gh-aw-estc-docs-patrol-external.lock.yml:
- Around line 1397-1402: The download step with id download-agent-output uses
name: agent but the upload used ${{ needs.activation.outputs.artifact_prefix
}}agent, so when a prefix exists the download silently fails; update the
download step to use the same computed name (set name: ${{
needs.activation.outputs.artifact_prefix }}agent) so it matches the upload, and
remove or set continue-on-error to false to ensure failures surface instead of
allowing downstream steps to run without the artifact.
In @.github/workflows/gh-aw-estc-newbie-contributor-patrol-external.lock.yml:
- Around line 1337-1343: The download step with id download-agent-output is
using a hardcoded name "agent" which mismatches the upload name that includes
the prefix from needs.activation.outputs.artifact_prefix, causing silent misses;
update the step to use the prefixed artifact name "${{
needs.activation.outputs.artifact_prefix }}agent" for the with.name field (keep
path /tmp/gh/), and retain or reconsider continue-on-error behavior so failures
aren't silently ignored if you want processing to stop on missing artifacts.
In @.github/workflows/gh-aw-framework-best-practices.lock.yml:
- Around line 1445-1457: The download step "download-agent-output" is requesting
the hardcoded artifact name "agent" while the upload uses the dynamic name "${{
needs.activation.outputs.artifact_prefix }}agent", so update the download step
to request the same dynamic artifact name (use
needs.activation.outputs.artifact_prefix concatenated with "agent") and keep the
subsequent GH_AW_AGENT_OUTPUT export logic (ensure the if condition still checks
the download step outcome and sets
GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json when the artifact is present) so
handlers below receive the correct agent_output.json; the referenced step id is
download-agent-output and the output variable to mirror is
needs.activation.outputs.artifact_prefix.
---
Outside diff comments:
In @.github/workflows/agent-deep-dive.lock.yml:
- Around line 649-700: The tool schema currently includes a branch for
"create_discussion" which must be removed so the workflow only exposes
"create_issue", "missing_tool", "missing_data", and "noop"; remove the
"create_discussion" entry from the tools array/schema (where "name":
"create_discussion" appears) and ensure any descriptions or prompts that
reference "create_discussion" are updated to require "create_issue" instead,
then update the original source markdown that generates this lockfile (the .md
that defines the tools) and recompile the lockfile with gh aw compile so the
generated .github/workflows/*.lock.yml no longer contains the create_discussion
branch.
In @.github/workflows/agent-efficiency.lock.yml:
- Around line 1056-1075: The uploaded artifacts (safeoutputs.jsonl and
agent_output.json) are created after the "Redact secrets in logs" step, so
either move the "Copy Safe Outputs" and "Ingest agent output" steps to run
before the "Redact secrets in logs" step or add a second redaction pass
immediately before the existing upload-artifact steps (the steps that upload
safeoutputs.jsonl/agent_output.json); update the "Copy Safe Outputs" and "Ingest
agent output" steps (and the collect_output action invocation) to ensure they
operate on the redacted files or invoke the same redaction logic used by "Redact
secrets in logs" so no secrets reach the upload-artifact steps.
In @.github/workflows/gh-aw-estc-pr-buildkite-detective.lock.yml:
- Around line 1357-1371: Add the GH_AW_FAILURE_REPORT_AS_ISSUE environment
variable to the source workflow file (gh-aw-estc-pr-buildkite-detective.md) so
it is compiled into the .lock.yml (match the name GH_AW_FAILURE_REPORT_AS_ISSUE
used in the handle_agent_failure step), and update the workflow permissions
block in that same .md file to include issues: write (in addition to the
existing read permissions) so the external handle_agent_failure.cjs can create
issues when the flag is enabled.
In @.github/workflows/gh-aw-framework-best-practices.lock.yml:
- Around line 149-159: The cross-repo hint currently checks failure() and
references steps.resolve-host-repo.outputs.target_repo which may not exist;
change the conditional to explicitly gate the "Cross-repo setup guidance" job on
the validate-secret step failing (e.g., if: steps.validate-secret.outcome ==
'failure') and then only compare against a resolved caller/host repo output that
is actually defined (ensure steps.resolve-host-repo exists in the source
workflow or use the correct step id that produces a target_repo output); update
the source .md workflow that generates this .lock.yml (not the compiled file) so
the compiled action includes the corrected if condition and valid step
reference.
---
Major comments:
In @.github/workflows/gh-aw-breaking-change-detector.lock.yml:
- Around line 1387-1399: The conclusion job is downloading the hardcoded
artifact name "agent" (step id download-agent-output) instead of the prefixed
name used when uploading (${ { needs.activation.outputs.artifact_prefix }
}agent), so replace the literal artifact name with the same prefixed expression
and ensure the conditional that sets GH_AW_AGENT_OUTPUT uses the successful
download step (download-agent-output) as before; update the template/source
fragment that generates this workflow (the file that sets artifact_prefix and
safe_outputs) and re-run the generator (gh aw compile) so the compiled .lock.yml
emits the prefixed artifact name and GH_AW_AGENT_OUTPUT is correctly set when
the download succeeds.
In @.github/workflows/gh-aw-bug-hunter.lock.yml:
- Around line 1382-1394: The workflow currently downloads the hardcoded artifact
name "agent" in the "download-agent-output" step causing the prefixed upload
(from needs.activation.outputs.artifact_prefix) to be missed; update the
download step's artifact name to use the uploaded prefix (i.e. "${{
needs.activation.outputs.artifact_prefix }}agent") so the file lands in
/tmp/gh-aw/ and GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json is correctly
set; keep the step id "download-agent-output" and the subsequent env export to
ensure downstream checks (noop/missing-tool/failure handling) read the prefixed
agent_output.json.
In @.github/workflows/gh-aw-code-complexity-detector.lock.yml:
- Around line 1503-1515: The download step with id download-agent-output is
fetching the hardcoded artifact name "agent" which mismatches the uploaded
artifact name that uses needs.activation.outputs.artifact_prefix; update the
download action to request the prefixed artifact name (concatenate
needs.activation.outputs.artifact_prefix with "agent") so the artifact is found
whenever a prefix is present, and ensure the subsequent Setup agent output
environment variable step still writes
GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json into $GITHUB_ENV (and preserves
the same /tmp/gh-aw/ path) so the job's safe_outputs flow receives the persisted
agent output.
In @.github/workflows/gh-aw-code-quality-audit.lock.yml:
- Around line 153-158: The conditional that gates the "Cross-repo setup
guidance" uses a non-existent step output
steps.resolve-host-repo.outputs.target_repo, so the hint can appear for
unrelated failures; either restore the resolver step that produces
resolve-host-repo.outputs.target_repo (e.g., re-add the resolve-host-repo step
that sets target_repo) or change the if: expression to a valid context such as
comparing github.repository to the caller repository (for example use
github.repository != github.event.repository or another existing step/output) so
the hint only runs for true cross-repo invocations.
- Around line 1407-1416: The download step and GH_AW_AGENT_OUTPUT should use the
same prefixed artifact name as the upload; update the download-artifact `name`
value from the literal "agent" to use the activation output prefix (i.e. "${{
needs.activation.outputs.artifact_prefix }}agent") and ensure the path and
subsequent GH_AW_AGENT_OUTPUT value reference the matching file (e.g.
/tmp/gh-aw/${{ needs.activation.outputs.artifact_prefix }}agent_output.json) so
the find and echo lines point to the correct prefixed artifact file.
In @.github/workflows/gh-aw-create-pr-from-issue.lock.yml:
- Around line 1346-1352: The conclusion job currently downloads a literal
artifact name "agent" (job id download-agent-output) but the activation job
uploads "${{ needs.activation.outputs.artifact_prefix }}agent", so when a prefix
is set the artifact is never found; update the artifact name in the download
step to use the same prefixed variable (e.g. set the download step's name field
to use ${{ needs.activation.outputs.artifact_prefix }}agent) in the source
workflow/fragment that generates these .lock.yml files and then re-run gh aw
compile to regenerate the .github/workflows/*.lock.yml outputs so all workflows
(including the other updated workflows in this PR) use the prefixed artifact
name consistently.
In @.github/workflows/gh-aw-deep-research.lock.yml:
- Around line 1345-1348: The workflow downloads the artifact named "agent" but
the activation job uploads it as "${{ needs.activation.outputs.artifact_prefix
}}agent", which leaves GH_AW_AGENT_OUTPUT unset for noop/failure handlers;
update the download-artifact step to use the prefixed artifact name matching the
upload (i.e., replace the static "agent" name with the same expression used by
the uploader: ${ { needs.activation.outputs.artifact_prefix } }agent) so the
artifact path and GH_AW_AGENT_OUTPUT are correctly set for all handlers and
persisted via the safe outputs mechanism.
- Around line 162-167: The failure message block incorrectly references a
non-existent step name resolve-host-repo and directs callers to set
COPILOT_GITHUB_TOKEN even though this workflow validates GEMINI_API_KEY; update
the source fragment or .md that generates this lock file so the cross-repo
remediation uses the correct outputs and secret name: remove or replace
references to steps.resolve-host-repo.outputs.target_repo, and change the error
guidance to mention GEMINI_API_KEY (and the correct secret configuration steps
for workflow_call cross-repo), ensuring the generated .lock.yml contains the
corrected step name/condition and error messages.
In @.github/workflows/gh-aw-dependency-review.lock.yml:
- Around line 150-155: The "Cross-repo setup guidance" step is gated on an
undefined step output (resolve-host-repo), so the hint can appear for unrelated
failures; either add/restore the missing resolver step named resolve-host-repo
that emits outputs.target_repo before this step, or change the step's if
condition to use an existing context/step (for example an existing step output
or github.repository comparison) that actually indicates a cross-repo
invocation; update the "Cross-repo setup guidance" step's if to reference the
restored resolve-host-repo outputs.target_repo or a valid existing context so
the guidance only runs for true cross-repo cases.
- Around line 1434-1443: The download step and the subsequent "Setup agent
output environment variable" assume the artifact name is literally "agent", but
the upload used "${{ needs.activation.outputs.artifact_prefix }}agent"; update
the download step's artifact name (used by actions/download-artifact and
referenced by steps.download-agent-output) to mirror the upload by using "${{
needs.activation.outputs.artifact_prefix }}agent", and ensure the
GH_AW_AGENT_OUTPUT environment variable (written to $GITHUB_ENV in the "Setup
agent output environment variable" step) still points to
/tmp/gh-aw/agent_output.json so the rest of the workflow reads the correct file
when the prefixed artifact is downloaded.
In @.github/workflows/gh-aw-duplicate-issue-detector.lock.yml:
- Around line 1266-1269: Update the artifact download step that currently uses
name: agent to use the prefixed artifact name from the activation output (e.g.
name: ${{ needs.activation.outputs.artifact_prefix }}agent) so the "conclusion"
path receives the correct artifact (agent_output.json); locate the
actions/download-artifact usage block and change the name value accordingly in
the source fragment or markdown that generates this .lock.yml so all branches
(noop and failure) persist the prefixed agent artifact via outputs.
In @.github/workflows/gh-aw-estc-docs-pr-review.lock.yml:
- Around line 1479-1482: The workflow step currently downloads the unprefixed
artifact name "agent" but the activation job uploads a prefixed artifact using
needs.activation.outputs.artifact_prefix, so change the download artifact name
to use the same prefix (i.e. name: ${{ needs.activation.outputs.artifact_prefix
}}agent) and ensure subsequent steps that expect agent_output.json still read
from /tmp/gh-aw/; also update the source template/markdown or shared fragment
that generates this lockfile so the prefixed artifact name is used everywhere
(preserving the artifact_prefix reference) so noop and failure handlers receive
the correct artifact.
In @.github/workflows/gh-aw-estc-pr-buildkite-detective.lock.yml:
- Around line 1316-1328: The Download agent output artifact step (id:
download-agent-output) currently requests the static artifact name "agent" but
the uploader uses the dynamic prefix from
needs.activation.outputs.artifact_prefix; change the artifact name in that step
to "${{ needs.activation.outputs.artifact_prefix }}agent" so the correct
prefixed artifact is downloaded, keeping the rest of the step (path /tmp/gh-aw/
and the follow-up step that sets GH_AW_AGENT_OUTPUT) unchanged.
In @.github/workflows/gh-aw-information-architecture.lock.yml:
- Around line 1380-1392: The download step (id: download-agent-output) currently
requests the literal artifact name "agent" but the artifact is uploaded as "${{
needs.activation.outputs.artifact_prefix }}agent", so GH_AW_AGENT_OUTPUT never
gets set; update the download-agent-output step to use the prefixed artifact
name from needs.activation.outputs.artifact_prefix (so it downloads "${{
needs.activation.outputs.artifact_prefix }}agent") and keep the subsequent Setup
agent output environment variable step that writes
GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json to GITHUB_ENV unchanged so the
environment variable is reliably populated.
In @.github/workflows/gh-aw-internal-gemini-cli.lock.yml:
- Around line 1367-1376: The workflow downloads and sets GH_AW_AGENT_OUTPUT
using the hard-coded artifact name "agent", but the upload used the prefixed
name from needs.activation.outputs.artifact_prefix; update the download-and-env
step (where steps.download-agent-output is referenced and GH_AW_AGENT_OUTPUT is
set) to request the same prefixed artifact by replacing the artifact name
"agent" with "${{ needs.activation.outputs.artifact_prefix }}agent" so the
downloaded file path and the GH_AW_AGENT_OUTPUT value point to the actual
uploaded artifact.
- Around line 159-164: The failure hint references a nonexistent step id
resolve-host-repo and tells callers to set COPILOT_GITHUB_TOKEN even though
activation never defines that step and the workflow validates GEMINI_API_KEY at
Line 154; update this branch to either use the correct cross-repo signal step id
(or add a resolve-host-repo step) or change the error text to reference
GEMINI_API_KEY and the actual condition used by activation so the hint matches
real validation logic (adjust the condition, step id, or message in the
Cross-repo setup guidance block accordingly).
---
Minor comments:
In @.github/workflows/agent-efficiency.lock.yml:
- Around line 1275-1287: The workflow currently sets GH_AW_AGENT_OUTPUT
unconditionally after the download-agent-output step, which only proves the
artifact exists, not that agent_output.json exists inside it; update the
download-agent-output block (step id download-agent-output) to check that
/tmp/gh-aw/agent_output.json actually exists before echoing GH_AW_AGENT_OUTPUT
to $GITHUB_ENV, and update any consumer steps that rely on GH_AW_AGENT_OUTPUT to
first guard on the environment variable being set and that the file exists
(e.g., replace bare references to GH_AW_AGENT_OUTPUT with a conditional if:
env.GH_AW_AGENT_OUTPUT != '' && <file-exists-check> or an equivalent run-time
check) so consumers only proceed when the file is present.
In @.github/workflows/gh-aw-code-complexity-detector.lock.yml:
- Around line 164-169: The conditional uses
steps.resolve-host-repo.outputs.target_repo but no step with id
"resolve-host-repo" exists, so update the workflow source to either add/restore
a step with id: resolve-host-repo that sets outputs.target_repo (the resolver
that detects the calling repo) or change the if condition to rely on an existing
context/output (for example compare github.repository with
github.event.repository.full_name or another defined step output) so the
Cross-repo setup guidance step can reliably detect cross-repo calls; adjust the
source .md workflow/fragment used by gh aw compile accordingly.
In @.github/workflows/gh-aw-create-pr-from-issue.lock.yml:
- Around line 160-165: The "Cross-repo setup guidance" step references
steps.resolve-host-repo.outputs.target_repo but the step generating that output
is missing the id; locate the source template or .md that emits the step which
should produce outputs.target_repo (the step referenced as resolve-host-repo),
add the missing id: resolve-host-repo to that step so the outputs are
addressable, then recompile the workflow with gh aw compile to regenerate the
YAML; ensure the id name exactly matches the referenced symbol
steps.resolve-host-repo so the if condition works at runtime.
In @.github/workflows/gh-aw-estc-docs-patrol-external.lock.yml:
- Around line 153-158: The conditional references a non-existent step id
steps.resolve-host-repo.outputs.target_repo (resolve-host-repo) so the job will
fail; either restore the missing step that sets outputs.target_repo (e.g., add
back the resolve-host-repo step that determines target_repo) or change the
conditional to use an existing output/context (for example github.repository,
github.event.repository.full_name, or the actual step id that resolves the host
repo) so the if: expression points to a valid output; update the job to ensure
the step id and output name exactly match the conditional (resolve-host-repo and
outputs.target_repo) if you choose to restore it.
---
Nitpick comments:
In @.github/workflows/gh-aw-code-complexity-detector.md:
- Around line 143-146: Replace the fragile pipeline "printf '%s\n' "$FILES" |
xargs wc -l | sort -nr | head -200 | awk '{print $2}'" with a null-delimited
safe variant: produce null-separated filenames from the $FILES variable and use
xargs -0 when invoking wc -l (and adjust the awk expression to ignore the total
line), or alternatively iterate over each filename with a while IFS= read -r
loop and call wc -l "$f" to avoid xargs entirely; update the pipeline
accordingly wherever the original command appears so paths with spaces or
special characters are handled safely.
| - name: Download agent output artifact | ||
| id: download-agent-output | ||
| continue-on-error: true | ||
| uses: actions/download-artifact@70fc10c6e5e1ce46ad2ea6f2b72d43f7d47b13c3 # v8 | ||
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | ||
| with: | ||
| name: agent-output | ||
| path: /tmp/gh-aw/safeoutputs/ | ||
| name: agent | ||
| path: /tmp/gh-aw/ |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
# First, let's locate the workflow file
find . -name "gh-aw-agent-suggestions.lock.yml" -type fRepository: elastic/ai-github-actions
Length of output: 121
🏁 Script executed:
# Get the total lines in the file to understand its size
wc -l .github/workflows/gh-aw-agent-suggestions.lock.ymlRepository: elastic/ai-github-actions
Length of output: 124
🏁 Script executed:
# Check line 1216 to see the upload artifact
sed -n '1210,1225p' .github/workflows/gh-aw-agent-suggestions.lock.ymlRepository: elastic/ai-github-actions
Length of output: 678
🏁 Script executed:
# Check lines 1375-1381 to verify the download snippet
sed -n '1370,1385p' .github/workflows/gh-aw-agent-suggestions.lock.ymlRepository: elastic/ai-github-actions
Length of output: 684
🏁 Script executed:
# Search for all references to artifact_prefix to understand the pattern
rg "artifact_prefix" .github/workflows/gh-aw-agent-suggestions.lock.yml -B 2 -A 2Repository: elastic/ai-github-actions
Length of output: 1997
🏁 Script executed:
# Check what the activation outputs section defines for artifact_prefix
rg -A 10 "outputs:" .github/workflows/gh-aw-agent-suggestions.lock.yml | head -50Repository: elastic/ai-github-actions
Length of output: 2091
🏁 Script executed:
# Verify the conclusion job structure and its dependencies
sed -n '1330,1350p' .github/workflows/gh-aw-agent-suggestions.lock.ymlRepository: elastic/ai-github-actions
Length of output: 1052
🏁 Script executed:
# Check job definitions to confirm conclusion depends on activation
rg "^\s*conclusion:" -A 5 .github/workflows/gh-aw-agent-suggestions.lock.ymlRepository: elastic/ai-github-actions
Length of output: 203
Download artifact name must use the same prefix as the upload.
The conclusion job uploads the agent artifact with a prefixed name (${{ needs.activation.outputs.artifact_prefix }}agent at line 1216) but downloads it with a hardcoded unprefixed name (agent at line 1380). When the prefix is non-empty, the download silently fails due to continue-on-error: true, skipping any downstream processing that depends on the agent output.
Fix
- name: agent
+ name: ${{ needs.activation.outputs.artifact_prefix }}agent📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| - name: Download agent output artifact | |
| id: download-agent-output | |
| continue-on-error: true | |
| uses: actions/download-artifact@70fc10c6e5e1ce46ad2ea6f2b72d43f7d47b13c3 # v8 | |
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | |
| with: | |
| name: agent-output | |
| path: /tmp/gh-aw/safeoutputs/ | |
| name: agent | |
| path: /tmp/gh-aw/ | |
| - name: Download agent output artifact | |
| id: download-agent-output | |
| continue-on-error: true | |
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | |
| with: | |
| name: ${{ needs.activation.outputs.artifact_prefix }}agent | |
| path: /tmp/gh-aw/ |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/gh-aw-agent-suggestions.lock.yml around lines 1375 - 1381,
The download step "download-agent-output" uses a hardcoded artifact name "agent"
but the upload used the prefixed name from
needs.activation.outputs.artifact_prefix (the upload at the "conclusion" job
used `${{ needs.activation.outputs.artifact_prefix }}agent`), causing mismatched
artifact names; update the download step's with.name to use the same prefix
expression (combine needs.activation.outputs.artifact_prefix with "agent") so
the actions/download-artifact call requests the identical artifact name produced
by the upload.
| - name: Cross-repo setup guidance | ||
| if: failure() && steps.resolve-host-repo.outputs.target_repo != github.repository | ||
| run: | | ||
| echo "::error::COPILOT_GITHUB_TOKEN must be configured in the CALLER repository's secrets." | ||
| echo "::error::For cross-repo workflow_call, secrets must be set in the repository that triggers the workflow." | ||
| echo "::error::See: https://github.github.com/gh-aw/patterns/central-repo-ops/#cross-repo-setup" |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Check if resolve-host-repo step exists in the workflow or related fragments
rg -n 'id:\s*resolve-host-repo' .github/workflows/Repository: elastic/ai-github-actions
Length of output: 51
🏁 Script executed:
#!/bin/bash
# Read the workflow file to see all step IDs defined
cat -n .github/workflows/gh-aw-docs-drift.lock.yml | grep -E '(^\s*-\s+name:|id:)' | head -50Repository: elastic/ai-github-actions
Length of output: 1288
🏁 Script executed:
#!/bin/bash
# Also check if there are any reusable workflows or imported workflows
find .github -type f -name "*.yml" -o -name "*.yaml" | head -20Repository: elastic/ai-github-actions
Length of output: 1048
🏁 Script executed:
#!/bin/bash
# Read the specific section mentioned in the review
sed -n '150,175p' .github/workflows/gh-aw-docs-drift.lock.ymlRepository: elastic/ai-github-actions
Length of output: 1477
🏁 Script executed:
#!/bin/bash
# Check all jobs and their steps to find where resolve-host-repo should be
grep -n '^\s*jobs:\|^\s*-\s*name:\|^\s*id:' .github/workflows/gh-aw-docs-drift.lock.yml | grep -A 5 -B 5 'Cross-repo'Repository: elastic/ai-github-actions
Length of output: 569
🏁 Script executed:
#!/bin/bash
# Look for any reference to resolve-host-repo anywhere in the file
rg -n 'resolve-host-repo' .github/workflows/gh-aw-docs-drift.lock.ymlRepository: elastic/ai-github-actions
Length of output: 162
🏁 Script executed:
#!/bin/bash
# Check if resolve-host-repo is defined in any imported/called workflow
rg -l 'resolve-host-repo' .github/Repository: elastic/ai-github-actions
Length of output: 3182
🏁 Script executed:
#!/bin/bash
# Search for any pattern that might be creating outputs.target_repo
rg -n 'target_repo' .github/workflows/gh-aw-docs-drift.lock.ymlRepository: elastic/ai-github-actions
Length of output: 162
🏁 Script executed:
#!/bin/bash
# Check one of the workflows that contains resolve-host-repo to see how it's actually used
rg -A 3 -B 3 'id:\s*resolve-host-repo' .github/workflows/gh-aw-mention-in-issue-by-id.lock.yml | head -20Repository: elastic/ai-github-actions
Length of output: 51
🏁 Script executed:
#!/bin/bash
# Try a simpler search without regex for resolve-host-repo in other workflows
rg 'resolve-host-repo' .github/workflows/gh-aw-mention-in-issue-by-id.lock.ymlRepository: elastic/ai-github-actions
Length of output: 158
🏁 Script executed:
#!/bin/bash
# Check what the actual step definition looks like in gh-aw-docs-drift.lock.yml around line 159-162
sed -n '159,162p' .github/workflows/gh-aw-docs-drift.lock.yml | cat -ARepository: elastic/ai-github-actions
Length of output: 319
🏁 Script executed:
#!/bin/bash
# Check if there's a job that might call another workflow that defines resolve-host-repo
grep -n 'uses:.*workflow' .github/workflows/gh-aw-docs-drift.lock.ymlRepository: elastic/ai-github-actions
Length of output: 51
🏁 Script executed:
#!/bin/bash
# Search the entire gh-aw-docs-drift.lock.yml for all job and step definitions
grep -n '^\s*jobs:\|^\s*name:\|^\s*id:' .github/workflows/gh-aw-docs-drift.lock.yml | head -30Repository: elastic/ai-github-actions
Length of output: 1238
🏁 Script executed:
#!/bin/bash
# Check what job this step belongs to and see if it's the activation job or another job
sed -n '108,165p' .github/workflows/gh-aw-docs-drift.lock.yml | grep -E '^\s*(jobs:|[a-z_-]+:)' | head -10Repository: elastic/ai-github-actions
Length of output: 335
🏁 Script executed:
#!/bin/bash
# Check the pre_activation job to see if resolve-host-repo is defined there
grep -n 'pre_activation:' -A 100 .github/workflows/gh-aw-docs-drift.lock.yml | grep -E '(name:|id:)' | head -20Repository: elastic/ai-github-actions
Length of output: 661
🏁 Script executed:
#!/bin/bash
# Search for the exact job structure - find all top-level job definitions
grep -n '^\s\s[a-z_-]*:$' .github/workflows/gh-aw-docs-drift.lock.yml | head -20Repository: elastic/ai-github-actions
Length of output: 182
🏁 Script executed:
#!/bin/bash
# Get the pre_activation job definition to check for resolve-host-repo step
sed -n '1481,1504p' .github/workflows/gh-aw-docs-drift.lock.ymlRepository: elastic/ai-github-actions
Length of output: 1049
Remove or fix the resolve-host-repo step reference.
This step references steps.resolve-host-repo.outputs.target_repo, but no step with id: resolve-host-repo exists in this workflow. The condition will fail to evaluate properly, and this guidance will never run on cross-repo failures. Either define the missing step or revise the condition to use an available output.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/gh-aw-docs-drift.lock.yml around lines 159 - 164, The
condition uses steps.resolve-host-repo.outputs.target_repo but there is no step
with id "resolve-host-repo"; either add or rename a step to expose
outputs.target_repo (e.g., create a step with id: resolve-host-repo that sets
output "target_repo") or change the if condition to reference an existing
step/output (for example replace steps.resolve-host-repo.outputs.target_repo
with the correct output name from the step that resolves the host repo or
compare directly to github.repository). Update the workflow so the if:
expression points to a valid step id/output (resolve-host-repo or the actual
existing step id) to ensure the cross-repo guidance runs correctly.
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | ||
| with: | ||
| name: agent-output | ||
| path: /tmp/gh-aw/safeoutputs/ | ||
| name: agent | ||
| path: /tmp/gh-aw/ |
There was a problem hiding this comment.
Artifact name mismatch: missing prefix in conclusion job.
The agent job uploads the artifact with ${{ needs.activation.outputs.artifact_prefix }}agent (line 1241), but this download step uses hardcoded agent. The download will fail to find the artifact.
Since this is a .lock.yml auto-generated file, the fix should be applied to the source .md file and recompiled.
Expected fix (in generated output)
- name: Download agent output artifact
id: download-agent-output
continue-on-error: true
uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1
with:
- name: agent
+ name: ${{ needs.activation.outputs.artifact_prefix }}agent
path: /tmp/gh-aw/🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/gh-aw-docs-drift.lock.yml around lines 1403 - 1406, The
download step uses a hardcoded artifact name "agent" but the upload uses the
prefixed name `${{ needs.activation.outputs.artifact_prefix }}agent`, so update
the source .md that generates this workflow to use the same prefixed name when
downloading (i.e., reference `${{ needs.activation.outputs.artifact_prefix
}}agent` in the download-artifact `name` field instead of `agent`) so the
generated .lock.yml contains the matching artifact name; regenerate the workflow
after making the change.
| id: download-agent-output | ||
| continue-on-error: true | ||
| uses: actions/download-artifact@70fc10c6e5e1ce46ad2ea6f2b72d43f7d47b13c3 # v8 | ||
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | ||
| with: | ||
| name: agent-output | ||
| path: /tmp/gh-aw/safeoutputs/ | ||
| name: agent | ||
| path: /tmp/gh-aw/ |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
# First, let's get the file size and context around the relevant lines
wc -l .github/workflows/gh-aw-estc-docs-patrol-external.lock.ymlRepository: elastic/ai-github-actions
Length of output: 132
🏁 Script executed:
# Read around lines 1237 (upload artifact)
sed -n '1230,1245p' .github/workflows/gh-aw-estc-docs-patrol-external.lock.ymlRepository: elastic/ai-github-actions
Length of output: 712
🏁 Script executed:
# Read around lines 1397-1402 (download artifact)
sed -n '1390,1410p' .github/workflows/gh-aw-estc-docs-patrol-external.lock.ymlRepository: elastic/ai-github-actions
Length of output: 929
Download artifact name does not match upload artifact name when prefix is applied.
Line 1401 downloads agent, but line 1237 uploads ${{ needs.activation.outputs.artifact_prefix }}agent. When the prefix is non-empty, the download silently fails (due to continue-on-error: true on line 1398), causing downstream steps to run without the agent result.
Fix
- name: agent
+ name: ${{ needs.activation.outputs.artifact_prefix }}agent📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| id: download-agent-output | |
| continue-on-error: true | |
| uses: actions/download-artifact@70fc10c6e5e1ce46ad2ea6f2b72d43f7d47b13c3 # v8 | |
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | |
| with: | |
| name: agent-output | |
| path: /tmp/gh-aw/safeoutputs/ | |
| name: agent | |
| path: /tmp/gh-aw/ | |
| id: download-agent-output | |
| continue-on-error: true | |
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | |
| with: | |
| name: ${{ needs.activation.outputs.artifact_prefix }}agent | |
| path: /tmp/gh-aw/ |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/gh-aw-estc-docs-patrol-external.lock.yml around lines 1397
- 1402, The download step with id download-agent-output uses name: agent but the
upload used ${{ needs.activation.outputs.artifact_prefix }}agent, so when a
prefix exists the download silently fails; update the download step to use the
same computed name (set name: ${{ needs.activation.outputs.artifact_prefix
}}agent) so it matches the upload, and remove or set continue-on-error to false
to ensure failures surface instead of allowing downstream steps to run without
the artifact.
| - name: Download agent output artifact | ||
| id: download-agent-output | ||
| continue-on-error: true | ||
| uses: actions/download-artifact@70fc10c6e5e1ce46ad2ea6f2b72d43f7d47b13c3 # v8 | ||
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | ||
| with: | ||
| name: agent-output | ||
| path: /tmp/gh-aw/safeoutputs/ | ||
| name: agent | ||
| path: /tmp/gh-aw/ |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
wc -l .github/workflows/gh-aw-estc-newbie-contributor-patrol-external.lock.ymlRepository: elastic/ai-github-actions
Length of output: 146
🏁 Script executed:
# Check the download-artifact section (lines 1337-1343)
sed -n '1337,1343p' .github/workflows/gh-aw-estc-newbie-contributor-patrol-external.lock.ymlRepository: elastic/ai-github-actions
Length of output: 333
🏁 Script executed:
# Check the upload-artifact section around line 1178
sed -n '1170,1185p' .github/workflows/gh-aw-estc-newbie-contributor-patrol-external.lock.ymlRepository: elastic/ai-github-actions
Length of output: 697
🏁 Script executed:
# Search for all artifact uploads with "agent" in the name to understand the pattern
rg -n 'name:.*agent' .github/workflows/gh-aw-estc-newbie-contributor-patrol-external.lock.ymlRepository: elastic/ai-github-actions
Length of output: 739
Download action uses unprefix artifact name, causing silent misses when prefix is non-empty.
Line 1342 downloads agent, but line 1178 uploads ${{ needs.activation.outputs.artifact_prefix }}agent. When the prefix is non-empty, the download silently fails and skips processing the agent result due to continue-on-error: true on line 1339. Note that another download-artifact step at line 1478 correctly uses the prefixed name.
Fix
- name: agent
+ name: ${{ needs.activation.outputs.artifact_prefix }}agent📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| - name: Download agent output artifact | |
| id: download-agent-output | |
| continue-on-error: true | |
| uses: actions/download-artifact@70fc10c6e5e1ce46ad2ea6f2b72d43f7d47b13c3 # v8 | |
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | |
| with: | |
| name: agent-output | |
| path: /tmp/gh-aw/safeoutputs/ | |
| name: agent | |
| path: /tmp/gh-aw/ | |
| - name: Download agent output artifact | |
| id: download-agent-output | |
| continue-on-error: true | |
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | |
| with: | |
| name: ${{ needs.activation.outputs.artifact_prefix }}agent | |
| path: /tmp/gh-aw/ |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/gh-aw-estc-newbie-contributor-patrol-external.lock.yml
around lines 1337 - 1343, The download step with id download-agent-output is
using a hardcoded name "agent" which mismatches the upload name that includes
the prefix from needs.activation.outputs.artifact_prefix, causing silent misses;
update the step to use the prefixed artifact name "${{
needs.activation.outputs.artifact_prefix }}agent" for the with.name field (keep
path /tmp/gh/), and retain or reconsider continue-on-error behavior so failures
aren't silently ignored if you want processing to stop on missing artifacts.
| - name: Download agent output artifact | ||
| id: download-agent-output | ||
| continue-on-error: true | ||
| uses: actions/download-artifact@70fc10c6e5e1ce46ad2ea6f2b72d43f7d47b13c3 # v8 | ||
| uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 | ||
| with: | ||
| name: agent-output | ||
| path: /tmp/gh-aw/safeoutputs/ | ||
| name: agent | ||
| path: /tmp/gh-aw/ | ||
| - name: Setup agent output environment variable | ||
| if: steps.download-agent-output.outcome == 'success' | ||
| run: | | ||
| mkdir -p /tmp/gh-aw/safeoutputs/ | ||
| find "/tmp/gh-aw/safeoutputs/" -type f -print | ||
| echo "GH_AW_AGENT_OUTPUT=/tmp/gh-aw/safeoutputs/agent_output.json" >> "$GITHUB_ENV" | ||
| mkdir -p /tmp/gh-aw/ | ||
| find "/tmp/gh-aw/" -type f -print | ||
| echo "GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json" >> "$GITHUB_ENV" |
There was a problem hiding this comment.
Conclusion still downloads the pre-prefix agent artifact.
Line 1450 asks for agent, but Line 1286 now uploads ${{ needs.activation.outputs.artifact_prefix }}agent. Once the prefix is non-empty, download-agent-output misses the artifact, GH_AW_AGENT_OUTPUT is never set, and the noop/missing_tool/failure handlers below lose agent_output.json.
Suggested fix
- name: agent
+ name: ${{ needs.activation.outputs.artifact_prefix }}agent🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In @.github/workflows/gh-aw-framework-best-practices.lock.yml around lines 1445
- 1457, The download step "download-agent-output" is requesting the hardcoded
artifact name "agent" while the upload uses the dynamic name "${{
needs.activation.outputs.artifact_prefix }}agent", so update the download step
to request the same dynamic artifact name (use
needs.activation.outputs.artifact_prefix concatenated with "agent") and keep the
subsequent GH_AW_AGENT_OUTPUT export logic (ensure the if condition still checks
the download step outcome and sets
GH_AW_AGENT_OUTPUT=/tmp/gh-aw/agent_output.json when the artifact is present) so
handlers below receive the correct agent_output.json; the referenced step id is
download-agent-output and the output variable to mirror is
needs.activation.outputs.artifact_prefix.
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
Requesting changes: the detector table in docs/workflows/gh-agent-workflows.md now includes five broken relative links (lines 74, 76, 77, 78, 81). The referenced targets do not exist, so docs navigation will 404 for users.
Please update those links to existing pages:
gh-agent-workflows/bug-hunter.md→gh-agent-workflows/bugs.mdgh-agent-workflows/code-duplication-detector.md→gh-agent-workflows/code-duplication.mdgh-agent-workflows/docs-patrol.md→gh-agent-workflows/docs-patrol-overview.mdgh-agent-workflows/newbie-contributor-patrol.md→gh-agent-workflows/newbie-contributor.mdgh-agent-workflows/text-auditor.md→gh-agent-workflows/text-quality.md
What is this? | From workflow: Trigger PR Review
Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.
1 participant
Summary
This PR shifts continuous-improvement automation to a detector-first chaining model and removes now-redundant standalone fixer workflows.
code-simplifierreferences/artifacts withcode-complexity-detector.gh-aw-create-pr-from-issuechaining:bug-exterminatorcode-duplication-fixernewbie-contributor-fixertest-improvertext-beautifiercode-simplifiergh-awbuild.Behavior and config updates
trigger-bug-hunter.yml,trigger-code-duplication-detector.yml,trigger-newbie-contributor-patrol.yml, andtrigger-test-coverage-detector.ymlnow run with write permissions and include acreate_pr_from_issuefollow-up job whencreated_issue_numberis present.trigger-code-complexity-detector.ymlwith scheduled and manual triggers plus the same in-run create-PR chaining pattern.README.md,gh-agent-workflows/README.md,scripts/quick-setup.sh) to prefer detector + chain patterns over dedicated fixer workflows.Documentation updates
gh-aw-create-pr-from-issue.docs/upgrading.mdwith explicit migration guidance from removed fixer/simplifier workflows to detector chaining.The body of this PR is automatically managed by the Trigger Update PR Body workflow.