Show and tell: independent review probes with a JSON-derived gate

[kyy-logs]

Sharing a code-review skill that follows a similar division of labor to multi-agent workflows: independent workers produce separate records, and the coordinator derives the final result from those records rather than from a single narrative summary.

In this case the workers are four review probes — behavior, tests, repository rules, impact — and each writes its own JSON. The human report and merge gate are both derived from that JSON. It is static-first and read-only.

It's a standalone Claude Code / Codex skill, not a Claude Flow integration. I'd be curious whether the record contract is useful for people composing review steps into larger workflows.

Repo: https://github.com/kyy-logs/code-review-skill