cleanexpo/pi-dev-ops/claude
Claude Code and agent skill by cleanexpo/pi-dev-ops
- community
Checking that the SKILL.md file is still in the repository.
Correct patterns for Claude SDK usage, subprocess vs API mode selection, MCP SDK imports, and streaming API call conventions in the Pi-Dev-Ops codebase. Use when touching Claude SDK invocations in Pi-Dev-Ops, choosing between `claude -p` subprocess mode and API key mode (`ANALYSIS_MODE`, Railway deployments), importing `McpServer` from `@modelcontextprotocol/sdk` subpaths, looping over `ClaudeSDKClient.receive_response()`, or passing an AbortController `signal` to `client.messages.stream()`.
Install
npx vskill@latest install cleanexpo/pi-dev-ops/claudeSKILL.md
The same SKILL.md was published earlier as cleanexpo/pi-dev-ops/claude.
Claude SDK & Runtime Patterns
When to apply
Apply whenever touching Claude SDK invocations, MCP server imports, subprocess vs API mode selection, or streaming message handling.
Rules
- The TAO engine in
src/tao/is stub scaffolding — it is NOT used by the web server. All intelligence is delegated to Claude via the SDK (API key mode). Check theANALYSIS_MODEenv var: Railway deployments must use API mode, not CLI mode. claude -psubprocess mode requires the Claude Code CLI installed in PATH. In cloud environments (Railway), Claude Code is NOT available — cloud deployments must useANTHROPIC_API_KEY+ SDK mode exclusively.- MCP SDK subpath imports are required:
'@modelcontextprotocol/sdk/server/mcp.js'and'../server/stdio.js'. The top-level'@modelcontextprotocol/sdk'import does NOT exportMcpServerdirectly — importing from it will fail silently or throw at runtime. - When using
claude_agent_sdk.ClaudeSDKClient, use the publicasync for message in client.receive_response()loop — NOT the private_query.receive_messages(). Private API methods break on SDK upgrades. - The AbortController
signalmust be passed toclient.messages.stream()as the second options argument:client.messages.stream({...}, { signal })— not inside the first object. Placing it inside the first argument silently ignores the abort signal.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/claude/SKILL.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/cleanexpo/pi-dev-ops/claude)