jarvis-doctor
Skill18 starsChanged 31 days agoMITView on GitHub
Health check for a Jarvis that's already built: is it running, reachable, inside its spend cap, up to date above the security floor, and still locked to its owner? Read-only by default. EN — "is my Jarvis ok", "Jarvis not answering", "health check". ES — "mi Jarvis no contesta", "revisa mi Jarvis", "está bien mi Jarvis", "doctor".
- Reads credentials
1 other project uses a version of this.
Adopted by
Most recently taken up by
- Erionis/Jarvis-Litetheir own version, first seen Aug 31, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- Erionis/Jarvis-Lite★ 3same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- jarvis-addSkill
- jarvis-nextSkill
- jarvis-quickSkill
- jarvis-refreshSkill
- jarvis-startSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
Read its jarvis-spec.md (copied into the folder at handoff) and RUNBOOK.md.
- Distinctive: line 18
Run read-only checks and report a traffic light per row:
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- Copyright (c) 2026 Enrique Rocha (@soyenriquerocha) · tododeia.com · id er-c08d5d7d -->/jarvis-doctor
Load the jarvis-start skill's rules (never read secrets; ask before any change).
- Find the Jarvis: argument path, else
.jarvis/tasks.json → jarvis_dir, else ask. - Read its
jarvis-spec.md(copied into the folder at handoff) andRUNBOOK.md. - Run read-only checks and report a traffic light per row:
| Check | How | |||
| Process alive | runtime's status command from knowledge/runtimes/<id>.md § Verify (launchctl/systemctl/pm2/railway status) | |||
| Runtime health | claude doctor · hermes doctor / hermes --version · openclaw doctor · ollama list | |||
| Version ≥ security floor | compare installed version with the floor in the runtime file; floor broken = 🔴 | |||
| Channel reachable | ask the owner to send "ping"; check the log line | |||
| Allowlist intact | grep the channel config for wildcards / open policies (fail on any) | |||
| Secrets not in git | `git -C <dir> ls-files | grep -E '.env$ | auth.json | credentials'` must be empty |
| Recent errors | last 50 log lines, summarised in ≤3 lines (never paste raw logs) | |||
| Scheduled jobs | last run time of each job | |||
| Cost | remind where to look (provider dashboard, Railway usage); you can't see billing | |||
| Autopilot | last run, paused or not, today's cost against JARVIS_AUTOPILOT_DAILY_USD, blocked tasks, TASKS.md parse errors, rejects in the last 7 days (commands below). Paused is 🟡, never "fixed" by you: only the owner resumes | |||
| Learning | the three hooks present (Claude tracks), captures still pending after 7 days, active lessons out of 60, skills on probation, learned skills failing the lint, last learning.review | |||
| Media | whisper-cli and ffmpeg on PATH, the model file's SHA-256 equal to the runtime file's pin, yt-dlp --version no more than 30 days behind the latest release, no leftover jarvis-watch-* folders. Skip on cowork | |||
| Panel (if installed) | GET /api/autopilot (paused, stopped, last_run, cost_today, actions_today) and GET /api/tasks (errors, needs_ok) |
Commands for the last four rows, run inside the Jarvis folder (they print counts and ids, never message content):
# Autopilot (knowledge/capabilities/autonomy.md)
grep autopilot logs/jobs.log | tail -1 # last run
test -e STOP_AUTOPILOT && echo paused || echo running
awk -F'\t' -v d="$(date +%Y-%m-%d)" '$2 == "autopilot" && index($1, d) == 1 { v = $5; sub(/^cost_est=/, "", v); s += v } END { printf "%.4f\n", s + 0 }' logs/audit.tsv
grep -c 'state:blocked' TASKS.md
python3 -I "$ROOT/tests/autonomy/parse_tasks.py" TASKS.md # exit 0 = parses
since=$(date -v-7d +%Y-%m-%d 2>/dev/null || date -d '7 days ago' +%Y-%m-%d)
grep 'autopilot.reject' logs/audit.jsonl | jq -r --arg s "$since" 'select(.ts >= $s) | .reason' | sort | uniq -c
# Learning (knowledge/capabilities/learning.md § 9)
for f in bin/hooks/capture-correction.sh bin/hooks/learn-nudge.sh bin/hooks/protect-paths.sh; do test -x "$f" || echo "missing $f"; done
python3 -I -c 'import json,sys,datetime as d;t=d.datetime.now(d.timezone.utc)-d.timedelta(days=7);print(sum(1 for l in open(sys.argv[1]) if l.strip() and (r:=json.loads(l)).get("outcome")=="pending" and d.datetime.fromisoformat(r["ts"].replace("Z","+00:00"))<t))' memory/corrections.jsonl
sed -n '/^## Activas/,/^## /p' LESSONS.md | grep -c '^- L-' # out of 60
find skills .claude/skills -name .probation 2>/dev/null
for d in skills/*/; do grep -q 'created-by: agent' "$d/SKILL.md" 2>/dev/null && { bin/promote-skill.sh lint "$d" >/dev/null || echo "lint fails: $d"; }; done
grep '"learning.review"' logs/audit.jsonl | tail -1 | jq -r .ts
# Media (knowledge/capabilities/multimodal.md)
command -v whisper-cli ffmpeg yt-dlp
shasum -a 256 models/ggml-*.bin # compare with the runtime file's pin
yt-dlp --version; gh api repos/yt-dlp/yt-dlp/releases/latest --jq .tag_name 2>/dev/null
ls "${TMPDIR:-/tmp}" | grep -c '^jarvis-watch-'yt-dlp versions are dates (YYYY.MM.DD): more than 30 days behind the latest tag is 🟡, with
the update command from the runtime file. With no gh, skip the comparison and say so.
- For each 🔴 / 🟡: one-line cause + the exact fix. Apply fixes only after a yes. Never resume
the autopilot, end a skill's probation or edit
LESSONS.mdyourself: name the panel button or the terminal command the owner runs.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

