How do I install the Inspect Profile skill?

Run npx skillmds@latest add thettwe/inspect-profile 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 Inspect Profile safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Inspect Profile?

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 Inspect Profile free to use?

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

Who published Inspect Profile?

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

Inspect Profile by thettwe — Agent Skill

Name: Inspect Profile
Author: thettwe

Inspect Profile

Explain in plain English what a nyann profile enables — stack, branching, hooks, extras, conventions, documentation. TRIGGER when the user says "what does the nextjs-prototype profile do", "explain the python-cli profile", "show me what <name> enables", "what's in profile X", "describe the <name> profile", "inspect profile X", "/nyann:inspect-profile". Do NOT trigger on "show this repo's setup" — that's the `explain-state` skill (reads the repo, not a profile file). Do NOT trigger on "apply profile X to this repo" — that's `bootstrap-project`. Do NOT trigger on "save this setup as a profile" — that's `learn-profile`.

thettwe 83dc672 2.8 KB Updated Sep 17, 2026

File contents

inspect-profile

Read-only. Wraps bin/inspect-profile.sh. Loads a profile via load-profile.sh (user profiles shadow starters with the same name) and renders a human-readable summary.

1. Resolve the profile name

The user names it explicitly most of the time ("what does nextjs-prototype do"). Use that.
If the user is ambiguous ("what does my profile do"), ask — the skill doesn't infer from the repo. For "show this repo's setup", route to explain-state instead.
Names match the regex ^[a-z0-9][a-z0-9-]*$. Reject anything else with a clear message rather than handing a bad name to the backend.

2. Invoke

bin/inspect-profile.sh <name> [--user-root <dir>]

--user-root defaults to ~/.claude/nyann. Only override when the user says something like "check my team-installed profile at <path>".

3. Read the output back

The backend already formats sections (Profile, Stack, Branching, Hooks, Extras, Conventions, Documentation, GitHub integration). Your job is to relay those faithfully and add context when the user asks a follow-up. Don't re-summarize — the backend's blurbs for each hook (e.g. "ESLint runs on staged JS/TS") are authoritative.

When the user asks "is this the right profile for my repo", do NOT guess — suggest running explain-state (reads the repo) and comparing, or running doctor (reports drift against a chosen profile).

4. Handle "not found"

Exit code 2 means the profile isn't in user root or starters. The backend prints the available list; relay it and ask the user to pick one. Never invent profile names.

5. Handle shadow warnings

load-profile.sh warns on stderr when a user profile shadows a starter with the same name. Don't hide this — explicitly tell the user which version was loaded (user wins by design).

When to hand off

"Apply this profile to a new repo" → bootstrap-project with --profile <name>.
"Change something in this profile" → edit the JSON at ~/.claude/nyann/profiles/<name>.json; the skill doesn't mutate.
"Save my repo's current state as a profile named X" → learn-profile.
"Why doesn't my repo match this profile" → doctor against the named profile.

---name: inspect-profiledescription: Explain in plain English what a nyann profile enables — stack, branching, hooks, extras, conventions, documentation. TRIGGER when the user says "what does the nextjs-prototype profile do", "explain the python-cli profile", "show me what <name> enables", "what's in profile X", "describe the <name> profile", "inspect profile X", "/nyann:inspect-profile". Do NOT trigger on "show this repo's setup" — that's the `explain-state` skill (reads the repo, not a profile file). Do NOT trigger on "apply profile X to this repo" — that's `bootstrap-project`. Do NOT trigger on "save this setup as a profile" — that's `learn-profile`.---# inspect-profileRead-only. Wraps `bin/inspect-profile.sh`. Loads a profile via`load-profile.sh` (user profiles shadow starters with the same name)and renders a human-readable summary.## 1. Resolve the profile name- The user names it explicitly most of the time ("what does `nextjs-prototype` do"). Use that.- If the user is ambiguous ("what does my profile do"), ask — the skill doesn't infer from the repo. For "show this repo's setup", route to `explain-state` instead.- Names match the regex `^[a-z0-9][a-z0-9-]*$`. Reject anything else with a clear message rather than handing a bad name to the backend.## 2. Invoke```bin/inspect-profile.sh <name> [--user-root <dir>]````--user-root` defaults to `~/.claude/nyann`. Only override when theuser says something like "check my team-installed profile at`<path>`".## 3. Read the output backThe backend already formats sections (Profile, Stack, Branching,Hooks, Extras, Conventions, Documentation, GitHub integration). Yourjob is to relay those faithfully and add context when the user asksa follow-up. Don't re-summarize — the backend's blurbs for each hook(e.g. "ESLint runs on staged JS/TS") are authoritative.When the user asks "is this the right profile for my repo", do NOTguess — suggest running `explain-state` (reads the repo) andcomparing, or running `doctor` (reports drift against a chosenprofile).## 4. Handle "not found"Exit code 2 means the profile isn't in user root or starters. Thebackend prints the available list; relay it and ask the user topick one. Never invent profile names.## 5. Handle shadow warnings`load-profile.sh` warns on stderr when a user profile shadows astarter with the same name. Don't hide this — explicitly tell theuser which version was loaded (user wins by design).## When to hand off- "Apply this profile to a new repo" → `bootstrap-project` with `--profile <name>`.- "Change something in this profile" → edit the JSON at `~/.claude/nyann/profiles/<name>.json`; the skill doesn't mutate.- "Save my repo's current state as a profile named X" → `learn-profile`.- "Why doesn't my repo match this profile" → `doctor` against the named profile.

Frequently asked questions