claude-automation-recommender

What it is

Claude Automation Recommender is a Claude Code skill that analyzes codebase patterns and recommends Claude Code automations across all extensibility options: hooks, subagents, skills, plugins and MCP servers. It is meant for cases where a user asks for automation recommendations, wants to optimize their Claude Code setup, mentions improving Claude Code workflows, asks how to first set up Claude Code for a project, or wants to know which Claude Code features they should use. The skill is explicitly read-only. It analyzes the codebase and outputs recommendations but does not create or modify any files. Users implement the recommendations themselves or ask Claude separately to help build them. It declares the tools Read, Glob, Grep and Bash. The skill works in three phases. In Phase 1 (Codebase Analysis), it gathers project context with shell commands. These detect the project type through files such as package.json, pyproject.toml, Cargo.toml, go.mod and pom.xml. They also check dependencies like react, next, prisma, supabase, convex and stripe, look for existing Claude Code config (.claude/, CLAUDE.md), and inspect project structure. It captures key indicators: language/framework, frontend and backend stacks, database, external APIs, testing, CI/CD, issue tracking and docs patterns. Each indicator maps to the kinds of automations it informs. In Phase 2 (Generate Recommendations), it maps codebase signals to specific recommendations. MCP server examples include context7 for live documentation, Playwright, Supabase, Convex, Database, GitHub, Linear, AWS, Slack, Memory, Sentry and Docker MCPs. For skills, it covers existing ones like commit, frontend-design and feature-dev, plus custom skills to create such as api-doc, create-migration, gen-test, pr-check and release-notes. Hook examples include PostToolUse auto-format, auto-lint, type-check or run related tests, and PreToolUse to block .env or lock file edits. Subagent examples include code-reviewer, security-reviewer, api-documenter, performance-analyzer, ui-reviewer and test-writer. Plugin examples include anthropic-agent-skills, frontend-design and mcp-builder. A decision framework explains when each type fits, along with skill invocation control: disable-model-invocation: true for user-only, user-invocable: false for Claude-only, and omitting both so either can invoke. In Phase 3, it produces a markdown report with a Codebase Profile and sections for MCP Servers, Skills, Hooks and Subagents. The report includes only the 1-2 most valuable recommendations per category and skips irrelevant categories. If the user asks for a specific type, it focuses only on that type and gives 3-5 options. It is instructed to go beyond its reference lists by using web search for recommendations specific to the codebase's tools. It ends by telling users they can ask for more or request help implementing the recommendations. The skill also includes configuration tips: sharing .mcp.json with the team, the --mcp-debug flag, headless mode for CI, and permissions for hooks in .claude/settings.json.

Who it's for

• Developers setting up Claude Code for a project for the first time
• Users who want to optimize or improve their existing Claude Code workflows
• Teams wondering which Claude Code features (hooks, subagents, skills, plugins, MCP servers) fit their codebase
• Users looking for automation ideas specific to their frameworks, libraries and tooling

Requirements

Examples

show me more MCP server options

what other hooks would help?

# Pre-commit hook example
claude -p "fix lint errors in src/" --allowedTools Edit,Write

# CI pipeline with structured output
claude -p "<prompt>" --output-format stream-json | your_command

{
  "permissions": {
    "allow": ["Edit", "Write", "Bash(npm test:*)", "Bash(git commit:*)"]
  }
}