Development

Work out W-4 withholding elections (new job, job change, mid-year income jump, or an annual check-up) using the official IRS Tax Withholding Estimator, and translate the result into concrete W-4 form entries.

Apply Vercel's 70-rule React/Next.js performance optimization guide. Use when writing, reviewing, or refactoring React components or Next.js pages.

Complete browser automation with Playwright. Auto-detects dev servers, writes clean test scripts to /tmp. Test pages, fill forms, take screenshots, check responsive design, validate UX, test login flows, check links, automate any browser task. Use when user wants to test websites, automate browser interactions, validate web functionality, or perform any browser-based testing.

Encode invariants and business logic into the type system so illegal states are unrepresentable. Use when designing domain models, reviewing code for primitive obsession, building Value Objects, modeling state machines, or applying Parse-Don't-Validate. Covers Go, Python, and Java with concrete before/after examples.

Develop, debug, and manage Temporal applications across Python, TypeScript, Go, Java, .NET, Ruby, and Rust. Use when the user is building workflows, activities, or workers with a Temporal SDK, debugging issues like non-determinism errors, stuck workflows, or activity retries, using Temporal CLI, Temporal Server, or Temporal Cloud, or working with durable execution concepts like signals, queries, heartbeats, versioning, continue-as-new, child workflows, or saga patterns. Also use when the user...

Apply Spring Boot and Java coding standards when writing, reviewing,

Create a new Claude Agent Skill following Anthropic's best practices with prompt engineering guidance

Analyze text, commands, or other content to determine if it should be converted to a Claude Agent Skill

Phase 6 — Dynamic idiom review, architecture review, refactor pass, then correctness + test gate before shipping.

Research market rates, build negotiation strategy, and create counter-offer scripts

Profile Rust binaries and benchmarks using cargo-flamegraph, samply, perf, and heaptrack. Covers CPU flamegraphs, collapsed stacks for LLM analysis, memory profiling, criterion benchmark profiling, and interactive Firefox Profiler UI.

Diagnose Rust performance bottlenecks from profiling data and apply targeted optimizations. Covers the full cycle: measure → profile (see rust-profiling skill) → diagnose → fix → verify. Includes pattern-matched fixes for common flamegraph hotspots, allocation elimination, data layout, compiler settings, and parallelism tuning.

Exploit CPU, GPU, and multi-machine parallelism in Rust. Covers rayon (data parallelism), tokio (task/async parallelism), SIMD (auto-vectorization + wide crate), wgpu/cudarc/candle (GPU compute and ML), tonic/tarpc/quinn (multi-machine RPC), MPI/ractor/DataFusion (HPC and distributed), algorithm patterns (map-reduce, pipeline, work-stealing, SPMD), profiling parallel programs, and common anti-patterns. Companion to rust-profiling (collect data first) and rust-perf-tuning (single-core fixes).

Idiomatic review for Rust — async/Tokio daemon design, ports-and-adapters trait boundaries, and thiserror/anyhow error handling. Use when reviewing or writing Rust code that defines async trait "ports" (dependency-inversion boundaries), runs a single-threaded Tokio daemon (current_thread runtime + LocalSet/spawn_local), or crosses a wire/domain type boundary. Covers async fn in traits vs async-trait, Rc/RefCell vs Arc/Mutex, tokio::select! cancel-safety, blocking calls in async context, thise...

Hub skill for Rust development. Covers the full workflow from profiling → performance diagnosis → CPU/memory optimization → parallelism. Routes to the right specialist skill based on the symptom. Links all four Rust skills: rust-profiling, rust-perf-tuning, rust-memory-optimization, rust-parallel-processing.

Methodically decompose existing code and projects to extract architectural

Create targeted resume sections optimized for different experience levels and roles

Wine dynamic tracing tool skill. Use when tracing Win32 API calls from a Windows binary running under Wine: WINEDEBUG relay logging, WineDbg, GDB attach to Wine process, observing file access, network connections, registry reads, DLL load order, or SEH. Called from code-reverse-engineering-binary (Phase 4) and code-re-qt5 (Phase 3).

Radare2 (r2) disassembly and analysis tool skill. Use when running r2 commands, scripting with r2pipe, analyzing functions, navigating disassembly, recovering symbols, finding cross-references, patching bytes, or using r2frida for live instrumentation. Covers r2's command language: analysis (aa/aaa/avrr), info (ii/iS/iE/ir/iee/ic), search (/w /R /v4), type system (td/to/tp), r2pipe Python API, and radiff2. Called from code-reverse-engineering-binary and code-re-qt5 (Phase 1–2 deep analysis).

Kaitai Struct schema authoring tool skill. Use when writing .ksy specs for binary protocols or file formats, using the expression language (instances, bit fields, process transforms, valid assertions), compiling with ksc, validating with ksv, generating Python/Rust/C++ parsers, or debugging a failing parse. Called from code-reverse-engineering-binary (Phase 7) and code-re-qt5 (Phase 4–5). Current stable: Kaitai Struct v0.11 (2025).

Ghidra decompilation and analysis tool skill. Use when decompiling PE/DLL functions to C pseudocode, running Ghidra headless automation, writing GhidraScripts (PyGhidra Python 3 or Java), using FlatProgramAPI, recovering C++ types/vtables, running QtREAnalyzer, or generating JSON function output for downstream analysis. Called from code-reverse-engineering-binary (Phase 3) and code-re-qt5 (Phase 2).

Frida dynamic instrumentation tool skill. Use when hooking functions in a running process, capturing raw payload bytes from send/recv, intercepting function arguments or return values, using Stalker for coverage tracing, or CModule for high-performance hooks. NOTE: Direct Frida attachment to Wine processes crashes — use frida-gadget.dll injection instead (documented below). Called from code-reverse-engineering-binary (Phase 5) and code-re-qt5 (Phase 3c).

Post-mortem cycle that turns fixed bugs into structural enforcement. For each bug just fixed, classifies the root cause, finds the earliest possible detection point on the enforcement ladder (compile → lint → test → checklist), implements it, and verifies it would have caught the original bug. Language-agnostic. Goal is to eliminate the class, not just the instance.

Detect dead/unused code (imports, private members, and whole-project-unreachable declarations) with language-specific tooling and a false-positive checklist before deleting anything