Skill Claude CodeCodex
Verify that an AI-generated repository change matches the requested outcome, preserves unrelated behavior, and has adequate evidence before acceptance. Use for patches, commits, and pull requests.
Skill Claude CodeCodex
Map an unfamiliar software repository before changing it. Use when an AI coding agent must locate the right files, understand callers and conventions, identify verification commands, or avoid missing hidden interfaces and generated code.
Skill Claude CodeCodex
Turn an ambiguous software request into explicit, testable acceptance criteria before an AI coding agent changes a repository. Use for feature requests, bug reports, refactors, migrations, and tasks with hidden assumptions or unclear boundaries.
Skill Claude CodeCodex
Design and review timeouts, retries, backoff, circuit breakers, rate limits, bulkheads, and graceful degradation for unreliable dependencies. Use when an AI coding agent changes network calls, queues, workers, caches, or failure handling.
Skill Claude CodeCodex
Use Git safely while an AI coding agent investigates, edits, reviews, commits, rebases, merges, or publishes repository changes. Use when branch state, uncommitted work, history, conflicts, or destructive commands could affect recoverability or collaboration.
Skill Claude CodeCodex
Review an AI-generated coding change for application security risks at trust boundaries, including input handling, authorization, secrets, data exposure, injection, and unsafe failure behavior. Use before accepting security-relevant features, bug fixes, or refactors.
Skill Claude CodeCodex
Select, order, and combine skills for an AI coding-agent task without activating irrelevant, duplicate, or conflicting guidance. Use when several skills may apply, a workflow spans phases, or a skill library is being extended.
Skill Claude CodeCodex
Review a proposed or existing SKILL.md before publication for practical scope, precise triggers, evidence quality, counterexamples, safety, overlap, composability, and honest lifecycle status. Use when creating, revising, promoting, deprecating, or merging agent skills.
Skill Claude CodeCodex
Use linters, type checkers, security analyzers, and other static checks as evidence while changing a codebase. Use when fixing diagnostics, introducing types, changing analyzer configuration, or deciding whether a warning is real, irrelevant, or dangerous to suppress.
Skill Claude CodeCodex
Evaluate whether tests detect acceptance-relevant behavioral faults using boundaries, negative cases, differential checks, and mutation-oriented diagnostics. Use when tests are added, changed, weak, or contradicted by regressions.
Skill Claude CodeCodex
Review AI-agent tools and proposed calls for least privilege, bounded parameters, explicit authorization, isolation, and auditable high-impact actions. Use when tools, credentials, scopes, or permissions change.