Help us improve
Share bugs, ideas, or general feedback.
From skill-maker
Analyzes project codebase at [path] [focus-topic|--full] to create a skill encoding conventions, patterns, and implementation details for consistent Claude work.
npx claudepluginhub vcode-sh/vibe-tools --plugin skill-makerHow this command is triggered — by the user, by Claude, or both
Slash command
/skill-maker:from-projectThe summary Claude sees in its command listing — used to decide when to auto-load this command
# Create Skill from Project Analysis Analyze a user's project codebase and create a skill that teaches Claude the project's conventions, patterns, and implementation details. Future conversations with this skill loaded will have Claude work consistently with the project's established patterns. ## Input The user provides: `$ARGUMENTS` Parse the input: - **Path**: Project directory (default: current directory `.`) - **Focus topic** (optional): Specific implementation to focus on (e.g., "websocket", "auth", "api", "database", "resend", "stripe") - **`--full` flag** (optional): Scan the ent...
/skill-createAnalyzes local git history to detect coding patterns like commit conventions and workflows, generating SKILL.md files for Claude Code skills. Supports --commits, --output, --instincts flags.
/suggest-skillsSuggests new skills to create by analyzing project context for gaps in technologies, workflows, and common practices. Reports skill name, purpose, usefulness, complexity, and similar skills.
/skillsDiscovers and recommends relevant skills from 410+ library by analyzing project files, conversation context, and optional [category|search-term]. Outputs project-specific suggestions and categories.
/skill-createAnalyzes local git history to detect coding patterns like commit conventions, workflows, and architecture, generating SKILL.md files for Claude. Supports --commits, --output, --instincts flags.
/skill-createAnalyzes local git history to extract coding patterns like commit conventions, workflows, and architecture, generating SKILL.md files. Supports --commits, --output, --instincts options.
/initAnalyzes codebase at given path and generates CLAUDE.md with project conventions, architecture, build commands, and patterns.
Share bugs, ideas, or general feedback.
Analyze a user's project codebase and create a skill that teaches Claude the project's conventions, patterns, and implementation details. Future conversations with this skill loaded will have Claude work consistently with the project's established patterns.
The user provides: $ARGUMENTS
Parse the input:
.)--full flag (optional): Scan the entire project comprehensively instead of focusing on one topicExamples:
/skill-maker:from-project ./my-app websocket → Focused scan of WebSocket implementation/skill-maker:from-project ./my-app --full → Full project scan/skill-maker:from-project → Scan current directory, ask for focus topicIf no focus topic and no --full flag, ask:
"What should the skill focus on? Options:
Analyze a specific implementation within the project.
Read essential project files first:
package.json / requirements.txt / go.mod / Cargo.toml / composer.json (dependencies)tsconfig.json / pyproject.toml / config files (project settings).env.example / .env.local (environment variables, NOT .env itself)README.md (project description)Present project overview:
PROJECT CONTEXT: [project name]
════════════════════════════════
Language: [detected]
Framework: [detected]
Key dependencies: [top 10 relevant]
Structure: [brief tree]
Use Grep and Glob to find all files related to the focus topic:
For "websocket":
WebSocket, ws://, wss://, socket.io, onmessage, onopen, onclose, useWebSocket**/ws/**, **/socket/**, **/websocket/**For "auth":
auth, login, session, jwt, token, middleware, passport, clerk, nextauth, lucia**/auth/**, **/middleware/**, **/session/**For "api":
router, endpoint, handler, controller, route, fetch(, axios**/api/**, **/routes/**, **/controllers/**For "database"/"db":
prisma, drizzle, mongoose, sequelize, knex, sql, query, schema, migration**/db/**, **/models/**, **/schema/**, **/migrations/**For any package name (e.g., "resend", "stripe"):
Present discovery results:
FOCUSED SCAN: [topic] in [project]
═══════════════════════════════════
Files found: [X] files matching [topic]
Key files:
- [path/to/main-file.ts] (primary implementation)
- [path/to/config.ts] (configuration)
- [path/to/types.ts] (types/interfaces)
- [path/to/tests/] ([X] test files)
Read every discovered file and extract:
Architecture Patterns:
Code Patterns:
API Integration:
Error Handling:
Configuration:
Testing Patterns:
Type System (if TypeScript/typed language):
--full)Analyze the entire project comprehensively.
Same as Focused Step 1, but also:
Analyze the full project structure:
Analyze patterns that span the entire project:
For each significant dependency:
[project]-[topic]-patterns (e.g., myapp-websocket-patterns)[project]-conventions (e.g., myapp-conventions)---
name: [skill-name]
description: [Project] [topic] implementation patterns and conventions. Use when working on [project] [topic] code, creating new [topic] features, or debugging [topic] issues. Covers [pattern 1], [pattern 2], [pattern 3] following established project conventions.
metadata:
author: [user]
version: 1.0.0
source: project-analysis
source-project: [project path]
framework: [detected framework]
language: [detected language]
category: project-conventions
tags: [project, topic, framework, language]
---
For Focused Scan:
For Full Scan:
Focused:
references/code-examples.md - All extracted code patterns with file pathsreferences/api-reference.md - API surface for this feature (if applicable)references/testing-patterns.md - Test examples from the projectFull:
references/architecture.md - Detailed architecture documentationreferences/patterns.md - All code patterns with examplesreferences/api-reference.md - Full API documentationreferences/testing.md - Testing conventions and examplesreferences/configuration.md - All config options and env varsWhen including code examples from the project:
// From: src/lib/websocket.tsSKILL CREATED FROM PROJECT ANALYSIS
════════════════════════════════════
Source: [project path]
Mode: [Focused: topic] / [Full scan]
Skill: [skill-name]/
Analysis summary:
- [X] files analyzed
- [X] code patterns extracted
- [X] API endpoints documented
- [X] test patterns captured
- [X] configuration items documented
Files created:
SKILL.md [X] words
references/code-examples.md [X] words
references/api-reference.md [X] words
references/testing-patterns.md [X] words
Validation: [Grade]
How this skill helps:
- New [topic] features will follow established patterns
- Error handling stays consistent with project conventions
- Tests follow the existing testing approach
- Configuration and env vars are documented
Next steps:
/skill-maker:test [skill-path] → Verify triggers
/skill-maker:review [skill-path] → Quality audit