doctor

Étoiles14

Forks1

Mis à jour18 février 2026 à 02:17

Runs a diagnostic check on your learning environment — verifies setup, project structure, tools, and module-specific requirements.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Source

zainnab-sparq

zainnab-sparq/cc-self-train

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Métiers associésSOC

Basé sur la classification professionnelle SOC

Administrateurs de réseaux et de systèmes informatiquesProfessions informatiques et mathématiques·SOC 15-1244

SKILL.md

readonly

name	doctor
description	Runs a diagnostic check on your learning environment — verifies setup, project structure, tools, and module-specific requirements.

Environment Diagnostic

You are a diagnostic assistant. Check the user's learning environment and report issues clearly.

What to Check

Read CLAUDE.local.md from the cc-self-train root to determine the active project, language, OS, directory, and current module. If CLAUDE.local.md doesn't exist, report that no project is set up yet and suggest running /start.

Run these checks in order, reporting pass/fail for each:

Core Setup

CLAUDE.local.md exists — Is the progress tracker present in the cc-self-train root?
Workspace directory exists — Does workspace/<project-dir>/ exist?
Project CLAUDE.md exists — Does workspace/<project-dir>/CLAUDE.md exist?
Git initialized — Is workspace/<project-dir>/ a git repo? Run git -C workspace/<project-dir> status
Git has commits — Does the project have at least one commit?

Language Toolchain

Language tools available — Based on the language in CLAUDE.local.md, verify the appropriate toolchain:
- Python: python --version or python3 --version (3.10+)
- TypeScript/Node: node --version (18+), npm --version
- Go: go version (1.21+)
- Rust: rustc --version, cargo --version
- Canvas (HTML/CSS/JS): Skip — just needs a browser

Module-Specific Checks

Only check these if the user has reached or passed that module:

Module 3+ — Does .claude/rules/ exist in the workspace project with at least one rule file?
Module 4+ — Does .claude/skills/ exist in the workspace project with at least one SKILL.md?
Module 5+ — Does .claude/settings.json exist in the workspace project with hook configuration?
Module 6+ — Does .mcp.json exist in the workspace project?
Module 8+ — Does .claude/agents/ exist in the workspace project with at least one agent file?

Report Format

Present results as a checklist:

Use a checkmark for passing checks
Use an X for failing checks with a brief explanation of how to fix each one
Skip module-specific checks the user hasn't reached yet
End with a summary: "N/N checks passed" or "N issues found — here's how to fix them"

If everything passes, give an encouraging message like: "Your environment looks great! You're ready to continue with Module N."

Plus depuis ce dépôt

même dépôt

sync

zainnab-sparq/cc-self-train

Maintainer tool: audit curriculum against latest CC release, present update plan for approval, then execute. Does NOT commit.

2026-04-2414

security-review

zainnab-sparq/cc-self-train

Audit the current project's Claude Code configuration for known AI-assisted-development security footguns — hooks, skills, MCP servers, CLAUDE.md — and report findings with severity and remediation.

2026-04-1914

experience

zainnab-sparq/cc-self-train

Change your curriculum experience level (beginner / intermediate / advanced) mid-course. Takes effect immediately.

2026-04-1714

start

zainnab-sparq/cc-self-train

Interactive onboarding — helps new users pick a project, verify their dev environment, and scaffold their project directory. Run this when someone first opens Claude Code in this repo.

2026-04-1714

stuck

zainnab-sparq/cc-self-train

Re-explain the current curriculum step at a different level, slower, with different examples. Use whenever you feel lost — no judgment, no "try harder."

2026-04-1714

recap

zainnab-sparq/cc-self-train

Reviews what you've learned — summarizes CC features, concepts, and what you built in completed modules.

2026-02-1814

Environment Diagnostic

You are a diagnostic assistant. Check the user's learning environment and report issues clearly.

What to Check

Run these checks in order, reporting pass/fail for each:

Core Setup

CLAUDE.local.md exists — Is the progress tracker present in the cc-self-train root?

Workspace directory exists — Does workspace/<project-dir>/ exist?

Project CLAUDE.md exists — Does workspace/<project-dir>/CLAUDE.md exist?

Git initialized — Is workspace/<project-dir>/ a git repo? Run git -C workspace/<project-dir> status

Git has commits — Does the project have at least one commit?

Language Toolchain

Language tools available — Based on the language in CLAUDE.local.md, verify the appropriate toolchain:

Python: python --version or python3 --version (3.10+)
TypeScript/Node: node --version (18+), npm --version
Go: go version (1.21+)
Rust: rustc --version, cargo --version
Canvas (HTML/CSS/JS): Skip — just needs a browser

Module-Specific Checks

Only check these if the user has reached or passed that module:

Module 3+ — Does .claude/rules/ exist in the workspace project with at least one rule file?

Module 4+ — Does .claude/skills/ exist in the workspace project with at least one SKILL.md?

Module 5+ — Does .claude/settings.json exist in the workspace project with hook configuration?

Module 6+ — Does .mcp.json exist in the workspace project?

Module 8+ — Does .claude/agents/ exist in the workspace project with at least one agent file?

Report Format

Present results as a checklist:

Use a checkmark for passing checks

Use an X for failing checks with a brief explanation of how to fix each one

Skip module-specific checks the user hasn't reached yet

End with a summary: "N/N checks passed" or "N issues found — here's how to fix them"

If everything passes, give an encouraging message like: "Your environment looks great! You're ready to continue with Module N."