framework/engineering/skills/engineer-plugin-mcp/SKILL.md
Build AI IDE plugin MCP elements: stdio JSON-RPC servers, tool schemas, host wiring for Claude Code/Codex, event-name mapping, and validation.
npx skillsauth add korchasa/flowai engineer-plugin-mcpInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Use this skill for MCP tools inside an AI IDE plugin. Keep server implementation host-neutral and generate host-specific wiring separately.
Do not treat this skill as a full MCP protocol or host configuration specification. Protocol versions, SDK APIs, host config fields, approval policies, and validation commands can change. Before implementation, open the current official docs for the MCP protocol and every requested host, then cite or link the pages used in the output.
Fetch only the sections needed for the current task. For any host not listed, find its official MCP, plugin, extension, or connector docs first. If no official docs exist, say so and separate verified behavior from assumptions.
For implementation requests, produce:
For design-only requests, return the same shape as a concise design response without writing files or running smoke tests.
development
Use when the user asks to add TypeScript strict-mode code-style rules to AGENTS.md for a TypeScript project using strict mode. Do NOT trigger for Deno projects (use setup-agent-code-style-deno) or non-strict TS configurations.
development
Use when the user asks to add Deno/TypeScript code-style rules to AGENTS.md, or during initial Deno project setup when code-style guidelines need to be established. Do NOT trigger for non-Deno TypeScript projects (use setup-agent-code-style-strict), or for runtime-agnostic style advice.
testing
Use when the user provides a source (URL, file path, or free text) to save into the project's memex — a long-term knowledge bank for AI agents. Stores the raw source, extracts entities into cross-linked pages, runs a backlink audit, and updates the index and activity log. Do NOT trigger on casual reads; only when the intent is to persist a source into the memex.
development
Use when the user asks to audit a memex (long-term knowledge bank for AI agents) for orphans, dead SALP REFs, missing sections, contradictions, or index drift. Runs a deterministic structural check, layers LLM-judgement findings, optionally auto-fixes trivial issues with `--fix`. Do NOT trigger on general code linting.