Works with the AI Agents you already use
Git Triage
Prevention at the gate, rescue for the 3am disaster, hygiene for the quiet decay, and proof that the rescue actually works.
$9.99
Git Triage
Example session with this skill installed
git-triage sweep --workdir /tmp/drill
- Read your context and instructions
- Compiled the git triage
tier 2 (Growing), mode sweep
findings: 3
drill botched-rebase: pass
- secret_generic_token: 1
- large_file: 1
- stale_branch: 1
Connects securely to your tools. The creator never sees your data.
About this skill
Git Triage
A standing maintenance crew for git repositories, packaged as a SKILL.md-format skill for AI coding agents. Prevention at the gate, rescue for the 3am disaster, hygiene for the quiet decay, and proof that the rescue actually works.
What it does
- Complexity profiler - reads commits, contributors, files, dependencies, languages, CI, submodules and remotes, then picks a working tier: Lean, Growing, Structured, Enterprise-grade. Tiers choose sweep vs granular coverage; the safety floor never shrinks.
- Read-only detection - secrets (AWS keys, private keys, generic tokens), large files, tracked artifacts (.env, build dirs, .DS_Store, pyc, sqlite), stale branches, old stashes, unpushed work. Every finding prints with its one-command fix.
- Self-installing commit gate - install.sh stamps a pre-commit hook that blocks secrets and bloat at commit time, preserving any existing hook.
- Weekly fire drills - breaks a SANDBOX CLONE on purpose (bad reset, deleted branch, botched rebase, corrupt object, force-push gone wrong), runs the rescue tools against the wreckage, scores pass/fail into the repo's health state. The real repo is never touched.
- Release gate - pre-release fails when a blocking finding sits on the stage.
- Regression forensics - capture a baseline, then forensics names the candidate commits that introduced new findings.
- Repo diet - largest tracked files, big blobs in history, tracked junk, stale branches.
How an agent uses it
SKILL.md is the only always-loaded surface. conductor.md carries the trigger table and pipeline order; stations/ load on demand. Scripts do the measuring: scripts/gittriage.py (state spine, profiler, detection) and scripts/git-triage.py (CLI).
Safety floor
Detection is read-only. One destructive step = one plain-sentence proposal with an undo, and the undo ships before the step does. Fire drills run only on sandbox clones. Nothing touches remotes, rewrites shared history, or deletes without approval. Scripts make no network calls and read no credentials.
Fire drills cover five scripted scenarios (bad reset, deleted branch, botched rebase, corrupt object, force-push gone wrong); they do not invent new disasters. Destructive suggestions (branch deletes, stash drops) require human approval. CLI and agent workflow only; no GUI.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 day ago
Requires
Permissions
File Scopes
Detection is read-only. Destructive steps are proposed with an undo, never executed on their own. Scripts make no network calls and read no credentials.