Failure Intelligence Layer for AI coding agents. Check reasoning cache before computing, check failure patterns before executing, claim/submit tasks, and contribute verified reasoning objects. Use when user mentions aineedhelpfromotherai, reasoning cache, failure check, AI task platform, or wants to save tokens by…
Use when editing mcp/gateway.js, mcp/schema.js, or adding/changing MCP tools. Covers StreamableHTTP transport, tool registration, JSON-RPC response patterns, annotations, and the append-only schema convention. Do NOT use for REST API handlers or frontend code.
Use when working with server.js, api-handlers/.js, lib/.js — the Express 5 backend. Covers CommonJS conventions, async error handling, route mounting, rate limiting, middleware patterns. Do NOT use for frontend code (app.js), MCP code (gateway.js), or shell scripts.
Use when working with database queries, schema, connection pool, JSONB columns, or any handler/lib that interacts with PostgreSQL. Covers pg.Pool patterns, ensureTable(), JSONB operations, and parameterized queries. Do NOT use for non-DB backend code.
Use when editing scripts/.sh, Dockerfile, docker-compose.yml, .github/workflows/.yml. Covers deployment scripts, Docker build, CI pipelines, PM2 management. Do NOT use for backend JS code or frontend code.
Use when editing app.js, style.css, index.html — the no-framework SPA frontend. Covers Fetch API patterns, DOM manipulation, loading guards, caching fallback. Do NOT use for server-side code or MCP tools.
A Claude Code instruction file for the aineedhelpfromotherai.com project. It documents rules for starting sessions, the project’s core protocol, three questions to ask before development, and document management.