Add agentic workflow discussion-task-miner#232
Conversation
|
Merging to
|
|
Note Gemini is unable to generate a summary for this pull request due to the file types involved not being currently supported. |
abhimehro
deleted the
add-workflow-.github-workflows-discussion-task-miner.md-8484
branch
|
👋 Development Partner is reviewing this PR. Will provide feedback shortly. |
There was a problem hiding this comment.
Pull request overview
Adds a new gh-aw agentic workflow (“discussion-task-miner”) intended to periodically scan GitHub Discussions and open a small set of code-quality improvement issues, while persisting deduplication state in a repo-memory branch.
Changes:
- Introduces
.github/workflows/discussion-task-miner.mdworkflow manifest and agent instructions. - Adds the compiled generated workflow
.github/workflows/discussion-task-miner.lock.yml.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
.github/workflows/discussion-task-miner.md |
Defines schedule/permissions/tools/safe-outputs and detailed task-mining instructions. |
.github/workflows/discussion-task-miner.lock.yml |
Auto-generated compiled GitHub Actions workflow for execution. |
| ### Issue Creation | ||
| - Create **3-5 issues maximum** per run (respects rate limits) | ||
| - Each issue expires after 1 day if not addressed | ||
| - All issues tagged with `code-quality`, `automation`, `task-mining` |
There was a problem hiding this comment.
Markdown typo: there’s an extra trailing backtick in `task-mining```, which breaks inline code formatting.
| - jq * | ||
| - cat * | ||
| - date * |
There was a problem hiding this comment.
The bash tool allowlist is extremely permissive (jq *, cat *, date *), which effectively allows arbitrary arguments/paths (including system paths like /proc) when the agent runs. This weakens the workflow’s defense-in-depth and increases the risk of unintended data exposure in logs. Prefer an explicit allowlist of the exact commands/files needed (e.g., restrict cat/jq to the repo-memory files under memory/discussion-task-miner/ and remove wildcard variants).
| - jq * | |
| - cat * | |
| - date * | |
| - jq | |
| - cat memory/discussion-task-miner/*.json | |
| - cat memory/discussion-task-miner/*.md | |
| - date |
| cat memory/discussion-task-miner/processed-discussions.json 2>/dev/null || echo "[]" | ||
|
|
||
| # Load extracted tasks log | ||
| cat memory/discussion-task-miner/extracted-tasks.json 2>/dev/null || echo "[]" |
There was a problem hiding this comment.
The repo-memory file format is inconsistent in the instructions: Step 1 reads processed-discussions.json / extracted-tasks.json with a fallback of [] (implying a JSON array), but Step 6’s examples write JSON objects with last_run and nested arrays. Pick one schema and update both the read/write examples accordingly; otherwise the agent will likely produce/expect the wrong shape and fail deduplication logic.
| cat memory/discussion-task-miner/processed-discussions.json 2>/dev/null || echo "[]" | |
| # Load extracted tasks log | |
| cat memory/discussion-task-miner/extracted-tasks.json 2>/dev/null || echo "[]" | |
| cat memory/discussion-task-miner/processed-discussions.json 2>/dev/null || echo '{"last_run": null, "processed_discussions": []}' | |
| # Load extracted tasks log | |
| cat memory/discussion-task-miner/extracted-tasks.json 2>/dev/null || echo '{"last_run": null, "tasks": []}' |
Add agentic workflow discussion-task-miner