elliott-chong1/shepherd

This system is an automated agent management service that orchestrates autonomous ownership tasks. It has been refactored from a standalone Python script into a continuous, reactive Flow architecture. The system is triggered by webhooks and includes logic to debounce re-triggers, ensuring stable and controlled execution of the Shepherd agent.

The autonomous MR ownership agent, previously implemented as a standalone Python script (shepherd.py), has been removed and replaced by a new 'Shepherd' Flow. The webhook filter (webhook\_filter.py) has been updated to trigger this new Flow instead of the old script, and now includes a debounce label to manage re-triggering. The migration changes how the agent is invoked and managed, shifting from a one-shot execution model to a continuously reactive agent model.

scripts · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• CAI 39 → 41 (+1.6)
• Rubric changed (rubric-2026.08.19 → rubric-2026.09.15) — scores are not directly comparable.

• Code Health 100 → 100 (+0.0)
• Architecture 69 → 69 (+0.0)
• Maturity 22 → 22 (+0.0)
• Readiness 26 → 32 (+6.2)
• Security 99 → 100 (+0.3)

• Dependency hygiene not measured — dependency manifest found but not parsed for hygiene
• Medium CVE: [GHSA redacted] (requirements.txt)
• No exposed public API
• early-stage repository — too little history to judge knowledge freshness
• git history depth insufficient
• git history depth insufficient
• single-maintainer — knowledge-concentration (bus factor) risk

• Banned license: python-gitlab
• Medium CVE: [GHSA redacted] (requirements.txt)
• No dependency advisory monitoring
• Outdated: python-gitlab
• Outdated: requests

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

elliott-chong1/shepherd was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.