[autonomy-atomicity] Internal deep-dive workflow concurrency key cancels unrelated runs

[github-actions]

Autonomy / Atomicity Findings

1. Static concurrency group in deep-dive workflow cancels independent investigations

Category: Ordering dependency
File(s): .github/workflows/agent-deep-dive.md (lines 18-20, 22-29, 33-35)

Problem:
agent-deep-dive supports both scheduled execution and manual workflow_dispatch, including per-run inputs (target-workflow, run-count). However, it uses a single static concurrency key:

concurrency:
  group: $\{\{ github.workflow }}-agent-deep-dive
  cancel-in-progress: true

With this key, any newly started run cancels any in-flight run, even when investigating a different target workflow. In concurrent development, two maintainers/agents running separate deep dives can invalidate each other’s work, and scheduled runs can preempt manual investigations.

Suggested fix:
Scope the concurrency group to run intent (for example, include inputs.target-workflow and/or github.ref) so independent deep dives can run without collision. Keep cancellation behavior only for truly duplicate runs.

Suggested Actions

• Update .github/workflows/agent-deep-dive.md concurrency key to include differentiators such as inputs.target-workflow (fallback when empty) and github.ref, then recompile.
• Decide whether cancel-in-progress should remain true for manual dispatches, or be limited to scheduled duplicate runs.

Note

🔒 Integrity filter blocked 26 items

The following items were blocked because they don't meet the GitHub integrity level.

• #490 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #418 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #937 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #608 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #643 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #641 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #767 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #399 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #537 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #546 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #675 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #697 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #851 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #599 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #504 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• #461 search_pull_requests: has lower integrity than agent requires. The agent cannot read data with integrity below "approved".
• ... and 10 more items

To allow these resources, lower min-integrity in your GitHub frontmatter:

tools:
  github:
    min-integrity: approved  # merged | approved | unapproved | none

---

What is this? | From workflow: Trigger Autonomy Atomicity Analyzer

Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.