Skip to content

Prevent Linter Miner runs from completing without a terminal safe output #32531

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
pelikhan merged 2 commits into from
May 16, 2026
+4 −2
Merged

Prevent Linter Miner runs from completing without a terminal safe output #32531

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4d610c5
Initial plan
Copilot May 16, 2026
4b6cd9a
Improve linter-miner safe-output completion guardrails
Copilot May 16, 2026
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
6 changes: 4 additions & 2 deletions .github/workflows/linter-miner.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,7 +109,7 @@ Read `/tmp/gh-aw/agent/prior-linters.json` (preloaded from cache-memory) to load

## Step 2 — Mine Sources for Linter Ideas

Use the `discussion-miner` sub-agent and the `code-pattern-scanner` sub-agent **in parallel** to gather raw evidence. Collect their outputs.
Use the `discussion-miner` sub-agent and the `code-pattern-scanner` sub-agent **sequentially** to gather raw evidence. Run one, wait for completion, then run the other.

### Discussion mining (sub-agent output)

Expand Down Expand Up @@ -222,7 +222,9 @@ Call the `create-pull-request` safe output with:
- Always include a `URL` field in the `Analyzer` pointing to `https://github.com/github/gh-aw/tree/main/pkg/linters/<name>`.
- The `Doc` string must be a single sentence beginning with "reports".
- If the linter cannot be implemented (e.g., repeated compilation failures after two fix attempts), call `noop` explaining why, rather than ending without a safe output.
- Always call either `create_pull_request` or `noop` before finishing — never end without a safe output.
- Do not finish while any spawned sub-agent is still running.
- If any sub-agent step fails, stalls, or does not complete, call `noop` with a clear explanation instead of ending without output.
- Final turn requirement: call exactly one safe output (`create_pull_request` or `noop`) as your last action before finishing.

---

Expand Down