Skip to main content
Glama
hooks.md•4.23 kB
# Hooks Integration > **Experimental:** This feature is experimental and may change in future releases. Hooks let scaffold-mcp proactively suggest scaffolding methods when AI agents create new files. Instead of writing boilerplate manually, agents can use available templates. --- ## How It Works 1. **Before file creation**: When an agent tries to write a new file, the hook shows available scaffolding methods that match the file pattern 2. **After tool use**: The hook can provide follow-up guidance or track progress --- ## Claude Code Setup Add to `.claude/settings.json` or `.claude/settings.local.json`: ```json { "hooks": { "PreToolUse": [ { "matcher": "Write", "hooks": [ { "type": "command", "command": "npx @agiflowai/scaffold-mcp hook --type claude-code.preToolUse" } ] } ], "PostToolUse": [ { "matcher": "mcp__one-mcp__use_tool", "hooks": [ { "type": "command", "command": "npx @agiflowai/scaffold-mcp hook --type claude-code.postToolUse" } ] } ] } } ``` ### What Happens When Claude tries to write a file like `src/app/about/page.tsx`: 1. The PreToolUse hook intercepts the Write call 2. Detects the project's `sourceTemplate` from `project.json` 3. Finds scaffolding methods matching the file pattern 4. Shows available options to Claude: ``` 🎯 Scaffolding Methods Available Before writing new files, check if any of these scaffolding methods match your needs: **scaffold-nextjs-page** Create a new Next.js page with proper structure and metadata. Required: pageTitle, pageDescription, nextjsPagePath **Instructions:** 1. If one of these scaffold methods matches what you need, use the `use-scaffold-method` MCP tool 2. If none are relevant, proceed to write files directly ``` --- ## Gemini CLI Setup (WIP) > **Note:** Gemini CLI integration is a work in progress. Add to `~/.gemini/settings.json`: ```json { "hooks": { "beforeToolUse": [ { "command": "npx @agiflowai/scaffold-mcp hook --type gemini-cli.beforeToolUse", "matcher": "write" } ], "afterToolUse": [ { "command": "npx @agiflowai/scaffold-mcp hook --type gemini-cli.afterToolUse", "matcher": ".*" } ] } } ``` --- ## Hook Decisions | Decision | Claude Code | Gemini CLI | Behavior | |----------|-------------|------------|----------| | Allow | `allow` | `ALLOW` | Proceed with optional guidance message | | Deny | `deny` | `BLOCK` | Show message and block operation | | Skip | `skip` | - | Silently allow (no output) | | Warn | - | `WARN` | Show warning but allow | **Note:** The PreToolUse hook uses `deny` to show scaffolding options. This displays the message to Claude but doesn't actually block the Write operation—Claude can still proceed if no scaffold methods are relevant. --- ## Session Tracking Hooks automatically track executions per session: - Each file path is only checked once per session - Prevents repeated suggestions for the same file - Resets when the agent session ends --- ## CLI Commands ```bash # Claude Code hooks npx @agiflowai/scaffold-mcp hook --type claude-code.preToolUse npx @agiflowai/scaffold-mcp hook --type claude-code.postToolUse # Gemini CLI hooks (WIP) npx @agiflowai/scaffold-mcp hook --type gemini-cli.beforeToolUse npx @agiflowai/scaffold-mcp hook --type gemini-cli.afterToolUse ``` Hook commands read tool context from stdin and write responses to stdout (JSON format). --- ## Troubleshooting ### Hooks not triggering 1. Verify hook configuration is in the correct settings file 2. Check matcher pattern matches the tool (`Write` for PreToolUse) 3. Ensure `npx` can find `@agiflowai/scaffold-mcp` ### No scaffolding methods shown 1. Project must have `sourceTemplate` in `project.json` or `toolkit.yaml` 2. Template must have `scaffold.yaml` with defined features 3. Feature's `patterns` field must match the file path ### Wrong template detected 1. Check the project's `sourceTemplate` configuration 2. Verify you're in the correct project directory 3. Ensure the template exists in your templates folder

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/AgiFlow/aicode-toolkit'

If you have feedback or need assistance with the MCP directory API, please join our Discord server