Skill Claude CodeCodex
Use when the user asks Codex to finish an intentionally unfinished refactor after they edited code to show the desired direction, left notes, simplified APIs, or knowingly broke compilation/tests as a hint. Especially relevant for requests like "finish this rework", "complete my mid-refactor changes", or "I simplified…
Skill Claude CodeCodex
Diagnose and fix failing GitHub Actions CI by inspecting runs with the GitHub CLI, reproducing as much of the repository's formatting, linting, type-checking, build, and test surface locally as practical, then committing, pushing, and monitoring the exact pushed revision. Use when the user asks to fix CI, repair a…
Skill Claude CodeCodex
Use only when the user explicitly asks for an AI-agent handoff or plan review, including handoff-review to Claude/Codex, reviewer instructions, external review ingestion, feedback triage, or plan revision before implementation. Produces .agent-handoffs/ artifacts, preserves provenance, and requires reviewers to verify…
Skill Claude CodeCodex
Use when writing, modifying, debugging, testing, refactoring, optimizing, or reviewing Rust in Jeroen's preferred style, including APIs, domain models, parsers, ETL and data processing, stateful data structures, numerical code, optimization and search algorithms, and performance-sensitive kernels.
Skill Claude CodeCodex
Optimize an existing stable algorithm or hot path through reproducible whole-workload baselines, profiler-guided experiments, fast rejection, behavior and quality gates, and optional sub-agent acceptance. Use for iterative or overnight throughput or latency campaigns that must preserve conceptual behavior. Do not use…