When an agent says done, which files did it test?

Description and use

Squadra tracks which files were tested by agents, storing proofs and reviews in a local registry to increase merge reliability.

Problem

When multiple coding‑agent windows work on the same project, the chat messages that tests passed or a review completed do not indicate which files were actually checked, making it hard to trust the results before merging.

Outcome

Squadra enables traceable coordination of tests and reviews across multiple agents, ensuring that the final merge is based on current, verified evidence.

Workflow

• 1. Author agent (Claude) modifies code to fix a bug.
• 2. Agent runs agreed test command; test passes on the candidate files.
• 3. Reviewer agent (Codex) reviews the frozen candidate and records approval.
• 4. Someone edits the checked files, changing the candidate.
• 5. Tests and review are repeated for the new candidate; a final check ensures the candidate, proof and review match before closure.

Evidence

Squadra keeps assignments, acknowledgments, proof, reviews and decisions in a shared local registry.

Source

Its prove command records the check and file fingerprints.

Source

Before a task can close, the candidate, proof and review are checked together again.

Source

The toolkit requires Python 3.12+ and Git; Node.js is needed for the dashboard checks in the self‑test.

Source