Skill Claude CodeCodex
Review user-facing software changes for accessibility barriers across semantics, keyboard and other input, focus, structure, content, feedback, and assistive technology behavior. Use for web, mobile, desktop, CLI, and interactive UI changes; adapt the checks to the platform.
Skill Claude CodeCodex
Design and interpret evaluations of AI coding-agent behavior across realistic tasks, correctness, safety, usefulness, and maintainability. Use when creating benchmarks, comparing agents or skills, measuring regressions, or making claims about agent capability.
Skill Claude CodeCodex
Design, change, or review APIs and serialized interfaces while making compatibility, versioning, errors, and migration behavior explicit. Use for HTTP, RPC, CLI, event, library, configuration, and file-format interfaces used by other code or people.
Skill Claude CodeCodex
Evaluate and record software architecture choices using constraints, alternatives, evidence, decision boundaries, and reversible consequences. Use when an AI coding agent changes system boundaries, data ownership, integration patterns, concurrency models, or long-term structure.
Skill Claude CodeCodex
Restructure code without changing its observable contract. Use when an AI coding agent extracts modules, renames interfaces, reduces duplication, changes control flow, or modernizes implementation while behavior should remain stable.
Skill Claude CodeCodex
Review and improve build systems and CI workflows for reproducibility, least privilege, trustworthy artifacts, controlled inputs, and repeatable verification. Use when an AI coding agent changes build commands, workflow files, runners, artifact publication, release checks, or provenance.
Skill Claude CodeCodex
Design, debug, or review concurrent and asynchronous code for race freedom, state invariants, ordering, cancellation, resource ownership, and failure recovery. Use when an AI coding agent changes threads, tasks, queues, locks, caches, events, retries, or shared mutable state.
Skill Claude CodeCodex
Design and review application configuration, environment handling, and secret use without unsafe defaults, accidental disclosure, or environment drift. Use when an AI coding agent changes config files, environment variables, deployment settings, credentials, keys, or feature flags.
Skill Claude CodeCodex
Resolve conflicting claims, evidence, tests, or failures without forcing false consensus. Use when sources or observed behavior disagree and the coding agent must preserve uncertainty and decision boundaries.
Skill Claude CodeCodex
Design and review database, schema, storage, and data migrations for correctness, compatibility, observability, rollback limits, and safe recovery. Use when an AI coding agent changes persisted data, schemas, indexes, formats, or migration scripts.
Skill Claude CodeCodex
Identify dead code, duplicated literals, drift-prone declarations, and comments that restate implementation without deleting anything until usage and intent are verified. Use when an AI coding agent is asked to clean up unused code or redundant comments.
Skill Claude CodeCodex
Review dependency and supply-chain changes for provenance, vulnerabilities, integrity, execution behavior, and residual risk. Use when manifests, lockfiles, build scripts, or external artifacts change.
Skill Claude CodeCodex
Compare a candidate coding change with a trusted baseline or reference to detect observable behavioral divergence. Use when ordinary tests pass but semantic alignment remains uncertain.
Skill Claude CodeCodex
Apply an evidence-first philosophy while investigating, changing, or explaining a software repository. Use when correctness, uncertainty, scope, or the evidence behind a coding decision matters.
Skill Claude CodeCodex
Diagnose software failures by reproducing the symptom, separating observations from hypotheses, testing competing explanations, and making the smallest verified fix. Use for bugs, regressions, flaky behavior, failing builds, and unexplained runtime errors in AI coding-agent work.
Skill Claude CodeCodex
Turn verified coding requirements into a minimal, dependency-aware implementation plan with risks, verification evidence, and stopping conditions. Use before multi-file, cross-boundary, risky, or ambiguous AI coding-agent changes.
Skill Claude CodeCodex
Design and review software for multiple languages, locales, scripts, regions, calendars, currencies, units, and text directions. Use when an AI coding agent adds user-facing text, formatting, sorting, input, translation, or locale-sensitive behavior.
Skill Claude CodeCodex
Keep coding-agent skills and guidance current by tracking provenance, freshness, deprecation, supersession, contradictions, and review triggers. Use when evidence, dependencies, versions, or failures change.
Skill Claude CodeCodex
Add or review logs, metrics, traces, and diagnostic signals so AI coding agents can understand real behavior and failures without excessive cost, noise, or data exposure. Use when changing production behavior, debugging intermittent issues, or operating asynchronous and distributed code.
Skill Claude CodeCodex
Investigate and validate software performance changes using representative workloads, baselines, variance-aware measurements, and resource tradeoffs. Use when an AI coding agent is asked to make code faster, reduce latency or cost, increase throughput, or diagnose a performance regression.
Skill Claude CodeCodex
Review software changes that collect, use, store, transmit, log, infer, share, or delete personal or sensitive data. Use to reduce privacy risk through purpose limitation, minimization, access control, retention, redaction, and lifecycle-aware design.
Skill Claude CodeCodex
Protect AI coding agents from instructions hidden in repository files, issues, documents, web pages, tool output, model context, or memory. Use whenever an agent reads untrusted content, retrieves external context, summarizes data, or proposes an action based on mixed instructions and data.
Skill Claude CodeCodex
Design focused regression tests for AI-generated or human software changes from observable behavior, failure modes, and boundaries. Use when fixing a bug, adding behavior, changing an interface, or strengthening a weak test suite.
Skill Claude CodeCodex
Plan and review safe software releases with explicit preconditions, staged exposure, health signals, rollback boundaries, and recovery evidence. Use when an AI coding agent changes deployable behavior, migrations, configuration, or release automation.