A fast weekly vitals pass — tests, build, deps, TODOs — diffed against the last run so drift shows up as trend arrows.
Free & open · no signup · read-only — it ends by asking · nothing leaves your machine
You are working inside this repo. Mission: run a fast vitals pass — minutes, not an afternoon — and diff it against the last run so drift shows up as trend arrows, not surprises. Designed to re-run weekly.
Read-only checks. Your only write is the report file.
Phase 2 sweeps the codebase through every one of these, citing file and line for each finding.
pass/fail · count · runtime
succeeds? · time · bundle/artifact size
vulnerability count (audit) · outdated count
lint errors · type errors
TODO/FIXME count · total LOC · file count
commits since last check · most-churned files this period
One structured report at the repo root — or in reports/, if you keep one — the same shape every time, ready for a teammate — or the optional Studio — to act on.
Every Goal Prompt follows the same four steps, so results are consistent and repeatable — no matter which one you run.
Look for an existing HEALTH.md at the repo root and in reports/. If present, parse the previous scorecard and history…
Collect quickly, mostly via commands.
Trend arrow per vital.
Overwrite HEALTH.md at repo root (append to history, replace the rest).
Copy it and paste it into your agent inside the repo you want checked.
Install the goal plugin once — two commands — then just type /goal:health-check.
/plugin marketplace add GhostlyGawd/goal-prompts/plugin install goal@goal-promptsOr install only this Goal Prompt as /goal-health-check:
curl -fsSL https://goal-prompts.vercel.app/install | BRIEF=29 shLet an agent fetch it mid-conversation, or pull the raw Goal Prompt by URL.
https://goal-prompts.vercel.app/raw/29.mdEach run appends one dated row to HEALTH.md — drop the file on the Vitals Viewer and every vital becomes a sparkline with run-over-run deltas. And a ready-made GitHub Action runs this Goal Prompt every Monday and files the report as an issue, so the ritual survives without your memory: copy the workflow ↗
Nothing hidden — this is the whole Goal Prompt, verbatim. Read it in a minute, edit it, or copy it as-is.
# Goal: Recurring Health Check You are working inside this repo. Mission: run a fast vitals pass — minutes, not an afternoon — and diff it against the last run so drift shows up as trend arrows, not surprises. Designed to re-run weekly. Read-only checks. Your only write is the report file. ## Phase 1 — Load the baseline - Look for an existing HEALTH.md at the repo root and in `reports/`. If present, parse the previous scorecard and history table for diffing. - If absent, this run establishes the baseline; say so. ## Phase 2 — Take the vitals Collect quickly, mostly via commands; skip gracefully what doesn't apply: 1. **Tests** — pass/fail · count · runtime 2. **Build** — succeeds? · time · bundle/artifact size 3. **Dependencies** — vulnerability count (audit) · outdated count 4. **Static health** — lint errors · type errors 5. **Code load** — TODO/FIXME count · total LOC · file count 6. **Momentum** — commits since last check · most-churned files this period ## Phase 3 — Diff and judge - Trend arrow per vital: better / same / worse vs last run. - Flag regressions loudly — a vital that got worse is the headline. - One-line interpretation per flagged item: what likely caused it (check recent commits). ## Phase 4 — Report Overwrite `HEALTH.md` at repo root (append to history, replace the rest): 1. **Scorecard** — vital · value · trend arrow · note 2. **Regressions** — what got worse, the likely cause, the suggested response 3. **Three actions** — the most valuable things to do before next check 4. **History table** — one dated row per run, appended; never edit old rows Start the report with today's date. If `HEALTH.md` already exists, the baseline parsed in Phase 1 is the diff target — lead with what changed since. If `FIXLOG.md` exists, credit fixed findings to their logged commits in the diff. ## Rules - Speed is a feature: minutes, or it won't get re-run - Numbers over prose; the format must stay stable so diffs stay meaningful - Same vitals every run — consistency is what makes trends real - If a `reports/` directory exists at the repo root, write the report there instead of the root. - Before asking, present the top findings as a ranked list in plain words - Report only — end by asking whether to act on the regressions
reports/ directory exists at the repo root, write the report there instead of the root.Runs inside these playbooks — curated sequences you can launch with one paste: