Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
[instructions] Sync instruction files with release v0.89.21
Document the gh-aw audit --group flag and the safe-outputs.jobs
needs/artifacts per-job properties, which shipped after v0.89.21
but were missing from the agentic-workflow instruction files.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
  • Loading branch information
github-actions[bot] and claude authored Sep 24, 2026
commit 2546cc76393209f83dec0265f9ac226c8febe58a
1 change: 1 addition & 0 deletions .github/aw/cli-commands.md
Original file line number Diff line number Diff line change
Expand Up @@ -113,6 +113,7 @@ gh aw audit <run-id> # Audit a single run
gh aw audit <run-id> --json # JSON output
gh aw audit <base-id> <compare-id> # Diff two runs (regression detection)
gh aw audit <id1> <id2> <id3> --json # Multi-run diff
gh aw audit <id1> <id2> --group # Grouped findings: one entry per [run, code] with occurrence counts
```

**MCP equivalent**: `audit` tool (single run) / `audit-diff` tool (multi-run comparison)
Expand Down
4 changes: 4 additions & 0 deletions .github/aw/safe-outputs-runtime.md
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,10 @@ The `report-incomplete` safe-output is enabled by default and is distinct from `

Post-processing GitHub Actions jobs registered as MCP tools. Agents call the tool by its normalized name (dashes → underscores, e.g., `send_notification`). The job runs after the agent completes with access to `$GH_AW_AGENT_OUTPUT` (agent output JSON path). Use to integrate Slack, Discord, external APIs, databases, or any service requiring secrets. Import from shared files via `imports:`.

Two extra per-job properties:
- `needs:` (string or array) - sequence this job relative to other compiled jobs; valid targets are `agent`, `safe_outputs`, `detection`, `upload_assets`, `unlock`, or another job name (dashes normalized to underscores). Persists across recompiles, unlike hand-patching the lock file.
- `artifacts:` (array of strings) - filesystem paths/globs under `/tmp/gh-aw/` that this job's `inputs` reference so the compiler includes them in the unified agent artifact upload (paths the agent wrote at runtime are otherwise dropped). Directory entries must end with `/`; only secret-redaction-covered extensions (`.txt`, `.json`, `.log`, `.md`, `.mdx`, `.yml`, `.jsonl`, `.patch`) are persisted, and `..` or hidden-file segments are rejected at compile time.

- `scripts:` - Inline JavaScript handlers running inside the safe-outputs job handler loop

```yaml
Expand Down