Skill Claude CodeCodex
Compare, score, synthesize, retest, and plateau-stop candidate versions with source-grounded identity, target maps, calibrated objectives, broad category buckets with explicit coverage lenses, validation-tier claim caps, evidence ledgers, recomputable score packets, observability markers, artifact checksums…
Skill Claude CodeCodex
Compare, score, improve, retest, and stop candidate-version work with source-grounded identity, target maps, calibrated objectives, broad category buckets with explicit coverage lenses, validation-tier claim caps, evidence ledgers, recomputable score packets, observability markers, artifact checksums, correction…
Skill Claude CodeCodex
Compare, score, improve, retest, and stop candidate-version work with source-grounded identity, target maps, calibrated objectives, broad category buckets with explicit coverage lenses, validation-tier claim caps, evidence ledgers, recomputable score packets, observability markers, artifact checksums, correction…
Skill Claude CodeCodex
Always-on governance skill for efficient and effective agent cognition and action. Use for reasoning, writing, research, coding, files, artifacts, current-information, safety-sensitive, or multi-step agentic tasks to scope the target, choose the cheapest sufficient thinking and behavior, verify evidence, respect…
Skill Claude CodeCodex
Always-on governance skill for efficient and effective agent cognition and action. Use for reasoning, writing, research, coding, files, artifacts, current-information, safety-sensitive, or multi-step agentic tasks to scope the target, choose the cheapest sufficient thinking and behavior, verify evidence, respect…
Skill Claude CodeCodex
Always-on governance skill for efficient and effective agent cognition and action. Use for reasoning, writing, research, coding, files, artifacts, current-information, safety-sensitive, or multi-step agentic tasks to scope the target, choose the cheapest sufficient thinking and behavior, verify evidence, respect…
Skill Claude CodeCodex
Always-on governance skill for efficient and effective agent cognition and action. Use for reasoning, writing, research, coding, files, artifacts, current-information, safety-sensitive, or multi-step agentic tasks to scope the target, choose the cheapest sufficient thinking and behavior, verify evidence, respect…
Skill Claude CodeCodex
Compress context while preserving task-critical state, exact artifacts, validation, language/tone constraints, and handoff continuity. Use for low-token, brief, compression, continuation, or long-thread state-loss risk.
Skill Claude CodeCodex
Compress and conserve conversation context by using terse responses, compact progress updates, explicit context pruning, compact code/content summaries, and concise handoff summaries. Use when the user asks to reduce token usage, avoid excessive context consumption, compress context, summarize current state, keep…
Skill Claude CodeCodex
Compress and conserve conversation context by using terse responses, compact progress updates, explicit context pruning, compact code/content summaries, and concise handoff summaries. Use when the user asks to reduce token usage, avoid excessive context consumption, compress context, summarize current state, keep…
Skill Claude CodeCodex
Core coding-agent mechanics — task execution, code editing, verification, communication, context management, and persistent memory. Invoke once at the start of any coding/multi-step agentic session to load the full mechanics (the lean CLAUDE.md keeps only the always-on governor + cross-cutting guards).
Skill Claude CodeCodex
Core always-on coding-agent behavior bundle — task execution, code editing, verification, communication, context management, and persistent memory. Load for every coding/agentic session; combined variant of the six individual ai-skills.
Skill Claude CodeCodex
Core coding-agent mechanics — task execution, code editing, verification, communication, context management, and persistent memory. Invoke once at the start of any coding/multi-step agentic session to load the full mechanics (the lean CLAUDE.md keeps only the always-on governor + cross-cutting guards).
Skill Claude CodeCodex
Analyze a GitHub issue and its related source-code repository to identify the likely failure boundary, infer the root cause, map the problem to relevant files, and produce practical code-fix suggestions with patch templates, regression tests, commit text, and PR-ready messaging.
Skill Claude CodeCodex
Token-efficient response shaping that preserves correctness, safety, evidence, exact tokens, language, CTA, and audit handoff.
Skill Claude CodeCodex
Evidence-grounded reviewer that turns a proposed change into a compact, source-located findings list, amendment plan, validation path, and audit handoff without overwriting originals.
Skill Claude CodeCodex
Standalone, token-efficient visual skill spanning the full agent arc — perceive, ground/act, extract, diagnose, link-to-code, verify, and trust — for images, screenshots, source-backed graphics, visual systems, and visual corpora. A compact always-load kernel routes to on-demand modules for modes/lenses, evidence…
Skill Claude CodeCodex
Choose the simplest representation that preserves what matters for the task, then expose the representation's assumptions and likely failure modes before relying on it. Use CHECKSUM, SEAM, GAP, and SCALE only when their structural preconditions actually apply. Prefer principles and targeted falsification over…
Skill Claude CodeCodex
Decide whether another artifact mutation is required to satisfy the agreed outcome. Convert the target into observable completion conditions, require evidence before authorizing another change, prefer the smallest sufficient mutation, and stop when no authorized blocker remains. Keep a minimal control loop because…
Skill Claude CodeCodex
Default first-line, domain-agnostic meta-reasoning control for choosing the right path. Use it to resolve an undecided choice, narrow an exploration, audit an invariant, investigate an owner-identified gap, or determine whether a validated owner can be used directly when no Pick exploration or audit is requested.…
Skill Claude CodeCodex
Senior-developer execution mode for the shortest SAFE working solution: direct patch, command, or smallest runnable code. Use for implementation, bug fixes, scripts, config changes, and minimal-output signals like "ponytail", "lazy mode", "minimal", "shortest path", "just the fix", "don't overthink it", "no…
Skill Claude CodeCodex
Minimal senior-developer execution skill for choosing the shortest safe working path without over-engineering.
Skill Claude CodeCodex
Minimal senior-developer execution skill for choosing the shortest safe working path without over-engineering.
Skill Claude CodeCodex
Minimal senior-developer execution skill: choose the shortest safe working path, execute it, and avoid process theater.