القائمة
Use MCP (Model Context Protocol) servers to give Claude Code access to external tools, APIs, databases, and services. Trigger on /mcp-integration, or when setting up external tool connections.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Control shell, filesystem, MCP, browser, external APIs, and risky tools. Enforce read-only-before-write, least-privilege, and verification discipline. Trigger on /agent-tool-safety, or when using shell commands, file operations, MCP tools, browser automation, or external APIs.
Final sanity check on AI-generated code before merging. Run the checklist to catch common failure modes. Trigger on /ai-coding-agent-review, or before finalizing any significant AI-generated change.
Review API design — endpoints, types, contracts, error handling, versioning, and consistency. Trigger on /api-design, or when designing or reviewing API specifications.
Implement backend API endpoints — routing, middleware, input validation, error handling, authentication, rate limiting, file upload. Trigger on /api-implementation, or when building endpoints, adding middleware, handling file uploads, or implementing API security.
Record architecture decisions as ADRs — compare options, document trade-offs, and track rationale. Trigger on /architecture-decision, or when choosing libraries, designing modules, splitting systems, or making significant technical decisions.
Make the agent continue work without asking unnecessary questions. Enforce the ask-less policy from CLAUDE.md. Trigger on /ask-less-execution, or when the agent is asking too many clarifying questions instead of discovering answers.
| name | mcp-integration |
| description | Use MCP (Model Context Protocol) servers to give Claude Code access to external tools, APIs, databases, and services. Trigger on /mcp-integration, or when setting up external tool connections. |
| user-invocable | true |
| allowed-tools | Read, Grep, Glob, Bash, Edit |
MCP (Model Context Protocol) lets Claude Code connect to external tools — databases, APIs, file systems, browsers, and custom services.
MCP servers are configured in .mcp.json at the project root:
{
"mcpServers": {
"my-db": {
"command": "node",
"args": ["path/to/mcp-server.js"],
"env": {
"DB_URL": "${DB_URL}"
}
}
}
}
Or in .claude/settings.json under mcpServers key.
| Purpose | Example |
|---|---|
| Database queries | Connect to PostgreSQL, SQLite, MySQL |
| API integration | GitHub, JIRA, Asana, Slack, Notion |
| File system | Advanced file operations, search |
| Browser automation | Web scraping, testing |
| Custom tools | Internal company APIs, CLI wrappers |
.env or environment variables for credentials, never hardcode.MCP tools can be called from skills. Reference them in allowed-tools:
---
name: my-skill
allowed-tools: [
"Read", "Grep", "Glob", "Bash",
"mcp__my-server__my-tool"
]
---
.mcp.json in version control (without secrets).This skill learns from experience. Mechanism:
Read .claude/experience/mcp-integration.md at skill start to benefit from past lessons.
Write to .claude/experience/mcp-integration.md after use if you learned something new.
Format: YYYY-MM-DD: <lesson> — one line per entry.
Evolve: If the same issue repeats 3+ times, update this SKILL.md itself.
This skill improves over time. After each use, append lessons learned.
Record format: YYYY-MM-DD: <lesson>
Read this at skill start to apply past lessons.