How do I install the Project Doctor skill?

Run npx skillmds@latest add wolverin0/project-doctor in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Project Doctor safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: executes scripts. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Project Doctor?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Project Doctor free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Project Doctor?

wolverin0 (@wolverin0) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Project Doctor by wolverin0 — Agent Skill

Name: Project Doctor
Rating: 4.2 (5 reviews)
Author: wolverin0

File contents

project-doctor

Run a health audit on the current project's orchestration scaffolding. Read-only.

When to use

Right after a /monitoring-setup rollout to confirm the generated monitoring.md references real files
When inheriting a project from a peer pane and you want a sanity check
Periodically (monthly) to catch drift - moved scripts, deleted skills, expired Telegram chat IDs
When omniclaude complains it can't find a script the frontmatter promised existed

How to invoke

The user types /project-doctor from any project root. You then:

Run the checker:

node "$HOME/.claude/skills/project-doctor/check.js"

(On Windows in bash: same command - $HOME resolves to /c/Users/<you>.)

The script prints a structured table to stdout: one row per check with verdict (PASS/WARN/FAIL/SKIP) and a remediation hint.
Read the output. Summarise to the user:
- All PASS: "All checks green - project orchestration is healthy."
- Any FAIL: list the failing checks + remediation. Offer to fix the ones you can fix in-session (e.g., update placeholder Telegram chat ID, regenerate stale graphify report, run /monitoring-setup if monitoring.md is missing).
- WARN only: list them but don't insist - the user decides.

Checks performed

Check	Severity	What it validates
`monitoring.md` exists	FAIL if missing	File present in project root
`monitoring.md` parses	FAIL on bad YAML	Frontmatter is valid YAML
`watchdog:` reference	WARN if path-shaped and missing	If frontmatter `watchdog:` looks like a script path, file must exist
`roadmap_file:` resolves	WARN	Referenced file exists
`state_checkpoint:` resolves	WARN	Referenced path exists (file or dir)
Telegram chat ID is real	WARN	Not equal to `CHANGEME_CHAT_ID` placeholder
Skills referenced exist	WARN	Each entry in frontmatter `skills:` resolves to `~/.claude/skills/<name>/SKILL.md`
`AGENTS.md` exists	WARN	File present in project root
`CLAUDE.md` exists	WARN	File present in project root
`graphify-out/GRAPH_REPORT.md` fresh	WARN if older than 30d	Mtime check
`.claude/settings.json` parses	FAIL on parse error	Valid JSON if file exists
Worktree status	INFO	Counts uncommitted files (never fails)

Remediation hints (offer when found)

Missing monitoring.md -> /monitoring-setup
Stale graphify report (>30d) -> /graphify to regenerate
Placeholder CHANGEME_CHAT_ID -> ask user for real Telegram chat ID, then Edit the frontmatter
Missing referenced skill -> tell the user the skill name and ask if it should be removed from the list or installed

Out of scope

Do not auto-fix anything that requires a decision (placeholder values, missing skills). Report and wait.

---name: project-doctordescription: Read-only audit for stale or broken project orchestration files, agent instructions, monitoring config, graphify output, and referenced skills.---# project-doctorRun a health audit on the current project's orchestration scaffolding. Read-only.## When to use- Right after a `/monitoring-setup` rollout to confirm the generated `monitoring.md` references real files- When inheriting a project from a peer pane and you want a sanity check- Periodically (monthly) to catch drift - moved scripts, deleted skills, expired Telegram chat IDs- When `omniclaude` complains it can't find a script the frontmatter promised existed## How to invokeThe user types `/project-doctor` from any project root. You then:1. Run the checker:```bashnode "$HOME/.claude/skills/project-doctor/check.js"```(On Windows in bash: same command - `$HOME` resolves to `/c/Users/<you>`.)2. The script prints a structured table to stdout: one row per check with verdict (`PASS`/`WARN`/`FAIL`/`SKIP`) and a remediation hint.3. Read the output. Summarise to the user: - All `PASS`: "All checks green - project orchestration is healthy." - Any `FAIL`: list the failing checks + remediation. Offer to fix the ones you can fix in-session (e.g., update placeholder Telegram chat ID, regenerate stale graphify report, run `/monitoring-setup` if `monitoring.md` is missing). - `WARN` only: list them but don't insist - the user decides.## Checks performed| Check | Severity | What it validates ||---|---|---|| `monitoring.md` exists | FAIL if missing | File present in project root || `monitoring.md` parses | FAIL on bad YAML | Frontmatter is valid YAML || `watchdog:` reference | WARN if path-shaped and missing | If frontmatter `watchdog:` looks like a script path, file must exist || `roadmap_file:` resolves | WARN | Referenced file exists || `state_checkpoint:` resolves | WARN | Referenced path exists (file or dir) || Telegram chat ID is real | WARN | Not equal to `CHANGEME_CHAT_ID` placeholder || Skills referenced exist | WARN | Each entry in frontmatter `skills:` resolves to `~/.claude/skills/<name>/SKILL.md` || `AGENTS.md` exists | WARN | File present in project root || `CLAUDE.md` exists | WARN | File present in project root || `graphify-out/GRAPH_REPORT.md` fresh | WARN if older than 30d | Mtime check || `.claude/settings.json` parses | FAIL on parse error | Valid JSON if file exists || Worktree status | INFO | Counts uncommitted files (never fails) |## Remediation hints (offer when found)- Missing `monitoring.md` -> `/monitoring-setup`- Stale graphify report (>30d) -> `/graphify` to regenerate- Placeholder `CHANGEME_CHAT_ID` -> ask user for real Telegram chat ID, then Edit the frontmatter- Missing referenced skill -> tell the user the skill name and ask if it should be removed from the list or installed## Out of scopeDo not auto-fix anything that requires a decision (placeholder values, missing skills). Report and wait.

Frequently asked questions

Project Doctor