Agents Reflection Skills
Meta-skills that let AI coding agents configure themselves
No more editing config files. Just tell your agent what you need.
Installation
Via Skills CLI (recommended):
claude install-skill CodeAlive-AI/agents-reflection-skills
Via plugin marketplace:
/plugin marketplace add https://github.com/CodeAlive-AI/agents-reflection-skills.git
/plugin install agents-reflection-skills@agents-reflection-skills
# Restart Claude Code for changes to take effect
What's Included
This plugin provides 7 skills that work across Claude Code, Codex CLI, Cursor, VS Code, Gemini CLI, and more:
| Skill | What It Does |
|---|---|
| mcp-management | Install and manage MCP servers across 10+ coding agents |
| hooks-management | Manage hooks and automation for Claude Code and Codex CLI |
| settings-management | Configure settings for Claude Code (JSON) and Codex CLI (TOML) |
| subagents-management | Create and manage subagents across Claude Code and Codex CLI |
| skills-management | Organize, discover, and share skills for coding agents |
| plugins-management | Package and publish your own plugins |
| optimizing-claude-code | Audit repos and optimize CLAUDE.md for agent work |
Lightweight: All 7 skill descriptions combined use less than 500 tokens in your context window.
Use Cases (skills)
mcp-management
Install and manage MCP servers across 10+ coding agents
Install to All Your Agents at Once
"Install the Postgres MCP server to Claude Code, Cursor, and VS Code"
Uses add-mcp to install MCP servers to 10+ coding agents (Claude Code, Cursor, VS Code, Claude Desktop, Gemini CLI, Codex, Goose, GitHub Copilot CLI, OpenCode, Zed) with a single command. Handles config format differences (JSON, YAML, TOML) automatically.
Connect to Your Database
"Connect Claude to my PostgreSQL database"
Installs the database MCP server, configures your connection string. Now you can query your data conversationally.
Install GitHub Integration
"Connect Claude to GitHub so it can create PRs and manage issues"
Installs the official GitHub MCP server. Claude can now create branches, PRs, and work with issues directly.
Multi-Agent Sync
"Make sure all my coding agents have the same MCP servers"
Configures MCP servers consistently across Claude Code, Cursor, VS Code, Gemini CLI, and other agents, handling each agent's config format and paths.
hooks-management
Manage hooks and automation for Claude Code and Codex CLI
Auto-Format Code
"Run Prettier on TypeScript files after every edit"
Adds a PostToolUse hook. Every file Claude touches gets formatted automatically — no more style inconsistencies.
Auto-Run Tests
"Run pytest whenever Claude edits Python files"
Adds a PostToolUse hook for *.py files. Instant feedback on whether changes broke anything.
Log All Commands
"Log every command Claude runs to an audit file"
Adds a PreToolUse hook that appends commands to ~/.claude/command-log.txt. Full audit trail.
Block Dangerous Commands
"Add hook that blocks global rm -rf commands"
Adds a PreToolUse hook that rejects destructive commands before they run. Protects against accidental data loss.
Require Confirmation
"Add hook that asks user permission for commands containing 'db reset'"
Adds a PreToolUse hook that pauses and requests confirmation for database resets or other sensitive operations.
settings-management
Configure settings for Claude Code (JSON) and Codex CLI (TOML)
Enable Sandbox Mode
"Turn on sandbox mode so Claude can work without asking permission for every command"
Enables native sandboxing — reduces permission prompts by 84% while keeping your system safe.
Block Access to Secrets
"Block Claude from reading .env files and anything in /secrets"
Adds deny rules to permissions. Sensitive files stay protected even if Claude tries to access them.
Switch Models
"Use Opus model for this project"