Creates Prowler Attack Paths openCypher queries using the Cartography schema as the source of truth for node labels, properties, and relationships. Also covers Prowler-specific additions (Internet node, ProwlerFinding, internal isolation labels) and $provider_uid scoping for predefined queries. Trigger: When creating or updating Attack Paths queries.
Updates the "Prowler at a Glance" table in README.md with accurate provider statistics. Trigger: When updating README.md provider stats, checks count, services count, compliance frameworks, or categories.
Start the OpenWork dev stack via Docker and verify real user flows via Chrome MCP. Triggers when user mentions: - "dev-up.sh" - "docker dev stack" - "verify in chrome mcp" - "test the real flow"
Expert knowledge for contributing to Litestream, a standalone disaster recovery tool for SQLite. Provides architectural understanding, code patterns, critical rules, and debugging procedures for WAL monitoring, LTX replication format, storage backend implementation, multi-level compaction, and SQLite page management. Use when working with Litestream source code, writing storage backends, debugging replication issues, implementing compaction logic, or handling SQLite WAL operations.
Analyze the latest AutoDeploy pipeline or a user-specified pipeline ID, inspect failed job logs, group similar failures into actionable root-cause buckets, and create at most one PR per bucket. Use when the user mentions pipeline IDs, failed jobs, GitLab logs, failure buckets, or opening PRs from CI failures.
Performance analysis coordination workflow. Guides profiling delegation, bottleneck classification (compute/memory/launch/communication/sync), and structured report generation. Use when the user asks to analyze performance, profile a workload, check MFU/SOL, or diagnose bottlenecks.
Analyze host/CPU overhead in TensorRT-LLM inference from nsys traces. Phase 1 (Detection): determine whether host overhead is the bottleneck via a binary YES/NO verdict with metric evidence (GPU idle ratio, host prep exposed ratio, per-phase breakdown). Phase 2 (Root Cause): isolate forward step iterations via allreduce kernel patterns, compare NVTX-instrumented host operations across versions, and pinpoint scheduling/request-management regressions. Usable standalone or as a sub-step of perf-analysis. Triggers: host overhead, inter-step gap, scheduling overhead, forward step isolation, nsys iteration analysis, NVTX breakdown, request management overhead, inference loop overhead, between-step gap, GPU idle, host bottleneck detection, host prep exposed.
Profiles and optimizes TensorRT-LLM host/CPU overhead using line_profiler (with nsys support planned). Runs iterative profile-analyze-optimize-validate rounds. Use when GPU utilization is low or optimizing PyExecutor throughput.
Analyze ncu (NVIDIA Nsight Compute) profiling output: SOL% bottleneck classification, roofline analysis, occupancy diagnosis, memory hierarchy analysis, warp stall analysis, metric interpretation, and programmatic .ncu-rep report analysis. NOT for kernel writing or code generation, Nsight Systems (nsys), host-side profiling, or system-level profiling.
Nsight Systems (nsys) CLI for system-level timeline profiling. Use when the user wants to run nsys profile, analyze .nsys-rep reports, use nsys stats/analyze/recipe commands, diagnose GPU idle time from timeline traces, or profile distributed training with NCCL overlap analysis. NOT for kernel-level metrics like SOL%, occupancy, or roofline (use perf-nsight-compute-analysis for ncu). NOT for writing or generating kernels. NOT for applying optimizations like CUDA Graphs.
Identify and eliminate host-device synchronizations in PyTorch code. Detects sync points (.item(), .cpu(), boolean indexing, torch.tensor on CUDA), classifies false vs true dependencies, provides sync-free alternatives. Triggers: sync-free, synchronization, .item(), .cpu(), host-device sync, eliminate syncs, CPU stall, non_blocking, set_sync_debug_mode, cudaStreamSynchronize, cudaEventSynchronize, remove syncs, async GPU.
Launch and test MetaMask Chrome extension with Playwright in headless sidepanel mode. Use for visual validation of UI changes, testing onboarding/unlock flows, confirmations, and capturing screenshots.
Run a browser-based UI review of the WordPress.com Help Center across multiple surfaces, looking for visual and behavioral issues. Use when asked to test the Help Center UI.
Run Rspack performance profiling on Linux using perf (with DWARF call stacks), generate perf.data, and analyze hotspots. Use when you need CPU-level bottlenecks, kernel symbol resolution, or repeatable profiling for rspack build/bench cases. Includes optional samply import with per-CPU threads for visualization, but primary analysis is perf-based.
Use sftrace, which is based on LLVM Xray instrumentation, to trace all Rust function calls. This can be used for performance analysis and troubleshooting.
Resolves eslint-disable suppression comments throughout the Bitwarden clients codebase by fixing the underlying issue. Use when the user asks to "fix FIXMEs", "fix eslint suppressions", "clean up eslint-disable-next-line", "resolve CL-764", "resolve CL-903", "fix OnPush eslint suppressions", "fix Signals eslint suppressions", or reduce linting suppressions.
Achieve aggressive v3 performance targets: 2.49x-7.47x Flash Attention speedup, 150x-12,500x search improvements, 50-75% memory reduction. Comprehensive benchmarking and optimization suite.
Run comprehensive worker system benchmarks and performance analysis
Tests in real browsers. Use when building or debugging anything that runs in a browser. Use when you need to inspect the DOM, capture console errors, analyze network requests, profile performance, or verify visual output with real runtime data via Chrome DevTools MCP.
Guides systematic root-cause debugging. Use when tests fail, builds break, behavior doesn't match expectations, or you encounter any unexpected error. Use when you need a systematic approach to finding and fixing the root cause rather than guessing.
Optimizes application performance. Use when performance requirements exist, when you suspect performance regressions, or when Core Web Vitals or load times need improvement. Use when profiling reveals bottlenecks that need fixing.
Run AddressSanitizer and UndefinedBehaviorSanitizer on the Z3 test suite to detect memory errors, undefined behavior, and leaks. Logs each finding to z3agent.db.
