V8 Cpu Profile Decoder Mcp

Turns V8 CPU profiles into token-efficient summaries your AI agent can actually read. Instead of choking on a 20MB .cpuprofile file, you get ranked hotspot lists, caller trees, source map resolution back to TypeScript, GC pressure analysis, before/after diffs, and async bottleneck detection. Six tools total: extract_hottest_functions pulls the top N offenders by self time, analyze_call_tree_path shows what's invoking your slow function, correlate_source_code maps compiled JS back to original .ts files, analyze_gc_pressure flags garbage collection overhead, diff_profiles compares two runs, and analyze_async_bottlenecks surfaces event loop saturation. Point it at a .cpuprofile from node --cpu-prof and ask which function is burning CPU or why your API is slow under load.