Skill Claude CodeCodex
Use when supplied imagery, footage, or sound needs source-faithful access artifacts authored: write functional alt text, time speaker-labeled captions with non-speech cues, and produce a complete transcript; mark decoration explicitly and stop if the asset cannot be perceived. Not for evaluating an interface against…
Skill Claude CodeCodex
Use when building or revising the system prompt for a product agent and you need an eval-backed boundary, tool-use, grounding, and output contract. Not for board-specialist adapters or lane capability projections, which follow model-lanes generators and controller policy.
Skill Claude CodeCodex
Use during an authorized adversarial assessment of an autonomous LLM application when attacker-controlled web content, MCP schemas, parameter descriptions, returned data, durable memory, or peer messages may cause credentialed actions: test injection, confused-deputy authority, approval enforcement, scope denial…
Skill Claude CodeCodex
Use when an adaptive game-audio design must become an engine-ready audio-event-map.json: define stable event and parameter IDs, cue transitions and cancellation, fallbacks, budgets, formats, and transition tests. Not for designing adaptive-audio behavior, rendering assets, or writing engine integration.
Skill Claude CodeCodex
Use during Project Phase 6 “Local deliver” when the verified repo still lacks one or more standard handoff files—README, CHANGELOG, LICENSE, or the agent-context file. Generate only approved missing files from what the work actually produced; never overwrite an existing file.
Skill Claude CodeCodex
Use to fact-check a finished deliverable's stated facts — its statistics, quotations, named sources, dates, and forecasts — by decomposing them into load-bearing factual claims and matching each to the exact source span that confirms or refutes it (Hard Rule 8 truth gate). Scope is whether what the text asserts as…
Skill Claude CodeCodex
Use before classifying a repository file or script as dead or proposing deletion: census static, dynamic, CI, hook, manifest, documentation, and host-scheduler callers; prove the search with positive and negative controls; classify reachability and resolve misleading mentions. Not for restructuring a live module graph.
Skill Claude CodeCodex
Use when reviewing a fixed diff against stated intent across correctness, blast radius, and maintainability, then returning severity-ranked, verified findings through bounded re-review iterations. The reviewer reports findings and a recommendation; the controller owns the accept/request-changes decision.
Skill Claude CodeCodex
Use when constructing a reusable palette for a visual system: choose a harmony, assign semantic roles and light/dark token values, account for brand and cultural meaning plus color-vision deficiency, and ensure interaction states never rely on hue alone. Not for measuring an existing interface against an accessibility…
Skill Claude CodeCodex
Use when specifying a voice or chat agent's dialogue before implementation: map intents and turns, write sample flows, define remembered context, cover no-input, no-match, ambiguity, correction, human-handoff recovery, and require explicit confirmation for consequential actions.
Skill Claude CodeCodex
Use when revising an existing draft for clarity, structure, and established voice while preserving its meaning and factual claims; produce a semantic diff and escalate uncertain voice choices to brand-voice. For new marketing copy, use brand-voice instead.
Skill Claude CodeCodex
Use when several files have been read independently but a conclusion depends on how control or data crosses them: build a typed, directed, line-anchored map of calls, imports, producer-consumer and config edges; identify cut points, blast radius, orphans, and traversal-backed conclusions. It maps arbitrary repository…
Skill Claude CodeCodex
Use when reviewing how a system handles sensitive or personal data: trace each data class through collection, storage, processing, sharing, retention, and deletion to surface privacy gaps, over-retention, or undisclosed sharing. Triggers on privacy review, DPIA, data-handling audit, or a schema/log that carries PII.
Skill Claude CodeCodex
Use when module, package, or build dependencies may contain cycles, especially load-order failures, broad rebuilds, or a planned extraction: derive the real graph, compute strongly connected components, rank cycle clusters, choose the minimal feedback-edge cut, and add an acyclicity guard.
Skill Claude CodeCodex
Use when a lockfile or vulnerability-scanner dump contains many dependency alerts and you must decide what actually needs action: derive the resolved transitive graph, merge advisories, prove vulnerable-symbol reachability and preconditions, rank intrinsic impact, assess maintenance risk, and name fixes or…
Skill Claude CodeCodex
Use when a detection engineer must author a Sigma, YARA, KQL, or SPL SIEM/EDR rule for a named TTP—pin the telemetry schema, add must-fire and must-not-fire fixtures, and replay representative history before any gated rollout. Not for incident triage or source-code defect scanning.
Skill Claude CodeCodex
Use when reviewing a particular code change against a large pre-existing Semgrep baseline: fix the base commit, choose rules by language and change shape, surface new or newly exposed source-to-sink paths, and report parse or coverage gaps. Route recurring false positives to semgrep-rule-author.
Skill Claude CodeCodex
Use when a proposed change claims to preserve or intentionally alter behavior and the delta must be proven between baseline and candidate—run both states, predict each observable shift, then compare return values, persistence, emitted events, logs, exit codes, wire formats, timing, errors, empty cases, and boundaries.…
Skill Claude CodeCodex
Use when numeric values cross arithmetic, API, storage, or time boundaries and a unit, scale, precision, rounding, signedness, or range mismatch could make the result wrong.
Skill Claude CodeCodex
Use when search results are too coarse to support a claim—sweep broadly to identify every candidate location, then open the most decisive hits around their definitions and callers before quoting or citing them.
Skill Claude CodeCodex
Use when a Figma file or frame is the authoritative implementation source and production UI must reproduce its tokens, component variants, auto-layout, interaction states, and responsive breakpoints through the project’s established primitives. This is source-to-code translation; post-build acceptance and inventing a…
Skill Claude CodeCodex
Use when reconstructing an incident from logs, images, records, or other artifacts and the chronology must preserve provenance—normalize clocks, hash and cite each source, label fact versus inference, and leave evidentiary gaps unknown.
Skill Claude CodeCodex
Use when shaping a new playable concept before downstream production—state the player fantasy and experience pillars, then define core/session/meta loops, verbs, goals, obstacles, resources, and win/lose/progression conditions as the authoritative upstream brief. It chooses what play is; later specialists quantify…
Skill Claude CodeCodex
Use when tuning an economy or ruleset calls for quantitative tradeoffs—model costs, payouts, rates, probabilities, player-proficiency assumptions, dominant choices, dead options, and runaway or stalling feedback, then state telemetry checks and provisional values. Deliver a numeric model, not an upstream concept brief…