Allow Step Name Alignment to read its generated manifest#51563
Conversation
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
|
@copilot this folder should be read/writeable my default in all engines |
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Addressed in |
PR Triage
Fixes #51547: scheduled Step Name Alignment workflow was failing sandbox checks. Change itself is a 1-line
|
There was a problem hiding this comment.
Pull request overview
Adds Claude filesystem access to /tmp/gh-aw/agent/ so workflows can read generated agent manifests.
Changes:
- Adds a default Claude
--add-dirargument. - Tests generated Claude arguments.
- Regenerates affected workflow lockfiles.
Show a summary per file
| File | Description |
|---|---|
pkg/workflow/claude_engine.go |
Adds the default directory argument. |
pkg/workflow/claude_engine_test.go |
Verifies generated arguments. |
.github/workflows/static-analysis-report.lock.yml |
Regenerates Claude commands. |
.github/workflows/smoke-github-claude.lock.yml |
Regenerates Claude commands. |
.github/workflows/smoke-claude-on-copilot.lock.yml |
Regenerates Claude commands. |
.github/workflows/smoke-agent-scoped-approved.lock.yml |
Regenerates Claude commands. |
.github/workflows/smoke-agent-public-none.lock.yml |
Regenerates Claude commands. |
.github/workflows/smoke-agent-public-approved.lock.yml |
Regenerates Claude commands. |
.github/workflows/smoke-agent-all-none.lock.yml |
Regenerates Claude commands. |
.github/workflows/smoke-agent-all-merged.lock.yml |
Regenerates Claude commands. |
.github/workflows/prompt-clustering-analysis.lock.yml |
Regenerates Claude commands. |
.github/workflows/lockfile-stats.lock.yml |
Regenerates Claude commands. |
.github/workflows/go-pattern-detector.lock.yml |
Regenerates Claude commands. |
.github/workflows/github-mcp-tools-report.lock.yml |
Regenerates Claude commands. |
.github/workflows/github-mcp-structural-analysis.lock.yml |
Regenerates Claude commands. |
.github/workflows/example-workflow-analyzer.lock.yml |
Regenerates Claude commands. |
.github/workflows/detection-analysis-report.lock.yml |
Regenerates Claude commands. |
.github/workflows/daily-rendering-scripts-verifier.lock.yml |
Regenerates Claude commands. |
.github/workflows/daily-elixir-credo-snippet-audit.lock.yml |
Regenerates Claude commands. |
.github/workflows/daily-choice-test.lock.yml |
Regenerates Claude commands. |
.github/workflows/daily-caveman-optimizer.lock.yml |
Regenerates Claude commands. |
.github/workflows/daily-astrostylelite-markdown-spellcheck.lock.yml |
Regenerates Claude commands. |
.github/workflows/copilot-session-insights.lock.yml |
Regenerates Claude commands. |
.github/workflows/copilot-agent-analysis.lock.yml |
Regenerates Claude commands. |
.github/workflows/cli-version-checker.lock.yml |
Regenerates Claude commands. |
.github/workflows/ci-doctor.lock.yml |
Regenerates Claude commands. |
.github/workflows/aw-failure-investigator.lock.yml |
Regenerates Claude commands. |
.github/workflows/avenger.lock.yml |
Regenerates Claude commands. |
.github/workflows/audit-workflows.lock.yml |
Regenerates Claude commands. |
.github/workflows/archivx-agentic-workflows-analyzer.lock.yml |
Regenerates Claude commands. |
.github/workflows/api-consumption-report.lock.yml |
Regenerates Claude commands. |
Review details
Tip
Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.
- Files reviewed: 61/61 changed files
- Comments generated: 1
- Review effort level: Balanced
| // and the allowed-tools string (reused for the comment annotation). | ||
| func (e *ClaudeEngine) buildClaudeCliArgs(workflowData *WorkflowData, toolsWithMountedCLIs map[string]any, logFile string) (claudeArgs []string, mcpConfigArg string, allowedTools string) { | ||
| claudeArgs = append(claudeArgs, "--print", "--no-chrome") | ||
| claudeArgs = append(claudeArgs, "--add-dir", constants.TmpGhAwAgentDir) |
|
@copilot Quick triage nudge for this PR. Please refresh the branch if GitHub can update it cleanly, investigate the current failed checks below, run the Failed checks: Please also review the latest Copilot review summary on the current head before handing back to maintainers. Branch refresh was requested. Run: https://github.com/github/gh-aw/actions/runs/31313728347
|
Co-authored-by: gh-aw-bot <259018956+gh-aw-bot@users.noreply.github.com>
Fixed in |
|
@copilot Quick triage nudge for this PR. Please address the open unresolved review thread and run the Open review context (newest first):
Branch refresh was requested. Run: https://github.com/github/gh-aw/actions/runs/31314812182
|
Verified: every compiled Claude workflow that invokes |
|
@copilot Quick triage nudge for this PR. Please address the open unresolved review thread and run the Open review context (newest first):
Branch refresh was requested. Run: https://github.com/github/gh-aw/actions/runs/31317059274
|
The review thread is fully addressed. The previous reply already documents the verification: every |
|
@copilot Quick triage nudge for this PR. Please resolve the remaining stale review thread on the current head, run the Open review context (newest first):
Branch refresh was requested. Run: https://github.com/github/gh-aw/actions/runs/31319713324
|
Step Name Alignment writes its input manifest under
/tmp/gh-aw/agent/, but Claude’s filesystem allowlist only included the repo checkout. The agent repeatedly hit sandbox denials when trying to inspectstep-alignment-input.json, causing scheduled runs to fail before analysis completed./tmp/gh-aw/agent/to Claude’s allowed directories viaengine.args.step-name-alignment.lock.ymlso the scheduled workflow passes the new--add-dirargument at runtime.Run: https://github.com/github/gh-aw/actions/runs/31313728347> Generated by 👨🍳 PR Sous Chef · gpt54 · 13.8 AIC · ⌖ 5.3 AIC · ⊞ 6.1K · ◷
Branch refresh requested by PR Sous Chef run https://github.com/github/gh-aw/actions/runs/31319713324> Generated by 👨🍳 PR Sous Chef · gpt54 · 15.9 AIC · ⌖ 5.45 AIC · ⊞ 6.1K · ◷