テスト・セキュリティのスキル

Audit whether a "current product focus receipt" (Decision/Now/Why-now/Evidence/Not-now/Cut-suspend/First-visible-proof/ Acceptance-gate/Kill-revisit-trigger/Owner/Review-date) and the work order it gates state a real, testable decision and a real agent launch — not planning-placeholder prose that sounds like an assignment but can't be proven. Use when deciding whether to launch an agent chain against a milestone or a stated "focus," when a work order lacks explicit input/output/owner/proof-gate, or when an acceptance gate would only prove a cached UI state instead of daemon truth. NOT for distinguishing roadmap items from energy-driven sidequests (legible-roadmap-with-sidequests), reviewing a PRD/prototype for product-completeness gaps like cold start or pricing (product-reality-reviewer), or accepting already-finished work against a personal acceptance bar (port-daddy-user-surrogate-pm-review).

日本語の概要は準備中です。原文の説明を表示しています。

Rigorous benchmarking of multi-agent coordination systems: experiment design, statistical analysis, human evaluation protocols, and reproducible reporting. NOT FOR ML model evaluation (use llm-evaluation-harness), A/B testing for web products, survey design, or general data science.

日本語の概要は準備中です。原文の説明を表示しています。

Use when introducing a feature flag system, choosing between release flags / experiment flags / operational flags / kill switches / permission flags, designing percentage rollouts, building kill-switch + circuit-breaker patterns, structuring evaluation contexts (user IDs, plan tier, region), or building flag-cleanup discipline. Triggers: LaunchDarkly / GrowthBook / Unleash / Flagsmith / OpenFeature, percentage ramp, sticky bucketing, kill switch wired to alerting, "we still have flags from 2 years ago", flag debt, OpenFeature provider, evaluation context, default value on provider failure. NOT for A/B test statistical analysis (separate skill), traffic-split routing at the edge, build-time bundler defines, or environment-based config.

日本語の概要は準備中です。原文の説明を表示しています。

Create and review Cloudflare Durable Objects. Use when building stateful coordination (chat rooms, multiplayer games, booking systems), implementing RPC methods, SQLite storage, alarms, WebSockets, or reviewing DO code for best practices. Covers Workers integration, wrangler config, and testing with Vitest. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

日本語の概要は準備中です。原文の説明を表示しています。

Architects the encryption, codec/transport, and premium-feature layer of a group video-calling product, on top of an already-chosen SFU topology. Covers E2EE patterns (Zoom leader-fan-out, Signal pairwise-fan-out, MLS/TreeKEM as in Discord's DAVE, Telegram's ledger consensus), the SFrame (RFC 9605) + WebRTC Insertable Streams mechanism letting an SFU relay encrypted frames it can't read, codec/transport choice (AV1/VP9 SVC vs. simulcast, Opus, congestion control, Media-over-QUIC, WHIP/WHEP), and which "premium" features (virtual backgrounds, live captions, recording, breakout rooms, spatial audio) survive E2EE vs. structurally can't. Use when adding E2EE to an SFU-based call, choosing a 2026 codec/SVC strategy, or deciding which features to disable under E2EE mode. NOT for call topology/platform choice (see webrtc-adhoc-video-chat, which this composes with). NOT for one-way broadcast (see managed-video-streaming-pipeline). NOT for the DM/group-chat text backend (see realtime-messaging-backend-architecture).

日本語の概要は準備中です。原文の説明を表示しています。

Audit designs, color palettes, and UI components for WCAG 2.1 accessibility compliance, providing specific fixes for contrast, touch targets, focus indicators, and screen reader support. NOT for color harmony aesthetics or design system creation.

日本語の概要は準備中です。原文の説明を表示しています。

Audit and maintain design system consistency, UX/UI patterns, color palettes, and design best practices across frontend applications

日本語の概要は準備中です。原文の説明を表示しています。

Designs and audits data retention and file compaction for an embedded SQLite datastore in a long-lived service — a unified retention registry (one declared policy per table), TTL / absolute-age / row-count-cap policies, a fail-loud coverage guard, WAL/checkpoint tuning, and auto_vacuum/incremental-vacuum so pruning actually shrinks the file. Use when a SQLite DB grows unbounded, a table has an index but no DELETE, a pruned DB never shrinks on disk, metrics are stored as raw events forever, or a new table needs bounding. Keywords retention policy, TTL, reaper, prune, VACUUM, auto_vacuum, incremental_vacuum, WAL checkpoint, freelist, DB bloat, row cap, per-tenant quota. NOT for Postgres/MySQL/server-DB partitioning or TTL, log-line rate-limiting/sampling, backup/replication, schema migrations, or query-performance indexing.

日本語の概要は準備中です。原文の説明を表示しています。

Audit or design server-side indexes, directories, reputations, or aggregates that are DERIVED from traffic a platform witnesses (chain heads, run verdicts, heartbeats, publish activity) so that consent gates the derivation, not merely the read: no shadow index over non-consenting principals, post-consent-only evidence windows, retention bounds and delete-propagation into derived rows, activity-oracle threat modeling, and audit-logged ranking discretion. Use when building a capability directory / whois from witnessed activity, a reputation ledger from settled outcomes, vitals aggregation across tenants, or any "the yellow pages can't lie because they index signatures" feature. Keywords: shadow index, derived index, consent gates derivation, activity oracle, phone book not a log, delist propagation, ranking weight audit, k-anonymous aggregate, subpoena surface. NOT for the consent-screen UX and scope-ladder mechanics themselves (use local-first-tenancy-boundary), reputation scoring math and Sybil economics (use agent-identity-continuity-reputation), or discovery-layer architecture L1/L2/L3 (use agent-discovery-directories-guilds).

日本語の概要は準備中です。原文の説明を表示しています。

Classify every destructive/gated action an agent body can attempt (git, filesystem, network, shell, github) into block/approve/allow tiers, and audit whether the pre-tool and post-tool governance backing that classification is real: blocked actions proven zero-side-effect by a negative fixture, every denial backed by a receipt and a transcript event, every block-tier denial paired with a safe alternative, and no unmanaged or same-UID body ever marked "contained." Use when building or reviewing Agent Harbor's C5 governance gate (destructive git blocker, approval request/result, denial receipts), gating a new tool surface before it can run destructive commands, or auditing whether an existing "destructive-action blocked" claim actually has evidence behind it. NOT for proving an existing sandbox boundary actually contains an adversary once an action is isolated (use sandboxed-adversarial-test-harness), deciding where in a DAG to place a human review node or designing the approval UX itself (use human-gate-designer), or designing the general work-receipt schema across an entire agent task (use agent-work-receipt-designer).

日本語の概要は準備中です。原文の説明を表示しています。

Managing third-party dependencies — version pinning, security auditing, license compliance, update workflows, lockfile management, supply chain security. Activate on "npm audit", "dependabot", "renovate", "pin versions", "dependency update", "supply chain", "license compliance", "lockfile", "security advisory", "typosquatting", "SBOM". NOT for internal monorepo package management (use monorepo-management) or publishing your own packages to npm/PyPI.

日本語の概要は準備中です。原文の説明を表示しています。

Runtime enforcement of file system boundaries and tool access restrictions. Blocks unauthorized operations and logs violations. Activate on 'enforce scope', 'access control', 'boundary enforcement', 'tool restrictions', 'runtime security'. NOT for validation (use dag-permission-validator) or isolation management (use dag-isolation-manager).

日本語の概要は準備中です。原文の説明を表示しています。

Validates versioned permission requests against a grant scope and records comparison evidence. Activate on 'validate permissions', 'permission check', 'inheritance validation', 'permission matrix', 'security validation'. NOT for runtime enforcement (use dag-scope-enforcer) or isolation management (use dag-isolation-manager).

日本語の概要は準備中です。原文の説明を表示しています。

Threat-models protocols where authority, scarce resources, bonds, bounties, escrow, reputation, ordering, adjudication, or settlement create strategic incentives. Use to test attack utility, conservation, collusion, Sybil resistance, oracle capture, griefing, custody, liquidity, and residual risk. NOT for choosing payment or bond amounts without evidence, smart-contract code audit, legal or employment classification, DeFi trading, admitting workers, or settling claims.

日本語の概要は準備中です。原文の説明を表示しています。

Build production computer vision pipelines for object detection, tracking, and video analysis. Handles drone footage, wildlife monitoring, and real-time detection. Supports YOLO, Detectron2, TensorFlow, PyTorch. Use for archaeological surveys, conservation, security. Activate on "object detection", "video analysis", "YOLO", "tracking", "drone footage". NOT for simple image filters, photo editing, or face recognition APIs.

日本語の概要は準備中です。原文の説明を表示しています。

Audit LLM token cost estimates against actual API usage. Activate on 'cost verification', 'token estimate accuracy', 'API cost audit', 'estimation variance'. NOT for pricing lookups, budget planning, or cost optimization strategies.

日本語の概要は準備中です。原文の説明を表示しています。

Build, debug, optimize, and deploy ComfyUI workflows for diffusion image, video, and audio generation. Activate on: ComfyUI workflow, ComfyUI custom node, ComfyUI API, ComfyUI Manager, KSampler, FLUX in ComfyUI, Wan 2.2 ComfyUI, Hunyuan video ComfyUI, LTX video ComfyUI, IPAdapter, ControlNet ComfyUI, Kijai wrapper, ComfyDeploy, RunComfy, ComfyUI security, GGUF quantization, TeaCache, Nunchaku, ACE-Step ComfyUI, F5-TTS ComfyUI, subgraphs ComfyUI. NOT for: A1111/Forge/Invoke (different UIs), training pipelines from scratch, non-diffusion ML models, or general image-API integrations (use generative-video-2026 / generative-music-audio / media-gen-deployment).

日本語の概要は準備中です。原文の説明を表示しています。

Detects and fixes color contrast violations using WCAG 2.1 guidelines and perceptual analysis. Expert in contrast ratio calculation, color blindness simulation, and providing accessible alternatives. Activate on "check contrast", "color accessibility", "WCAG audit", "readability check", "contrast ratio", "hard to read", "can't see text". NOT for general color theory (use color-theory-palette-harmony-expert), brand color selection (use web-design-expert), or non-visual accessibility (use ux-friction-analyzer).

日本語の概要は準備中です。原文の説明を表示しています。

Interprets GitHub Actions run status and logs, diagnoses CI failure patterns, and suggests targeted fixes. Handles dependency install failures, test timeouts, build OOM, flaky tests, and workflow misconfigurations. Activate on: 'CI failing', 'build broken', 'workflow error', 'GitHub Actions debug', 'flaky CI', 'pipeline timeout', 'CI red'. NOT for: writing new workflows from scratch (use github-actions-pipeline-builder), CI caching strategy (use ci-cache-optimizer), deployment orchestration (use devops-automator).

日本語の概要は準備中です。原文の説明を表示しています。

Diagnose when intuitive judgment, agent confidence, or expert routing can be trusted by classifying environment validity, feedback quality, and task-boundary fit. Use for confidence calibration, agent routing, expertise audits, and escalation design. NOT for deterministic implementation tasks, pure syntax debugging, or domains with explicit verifiable answers.

日本語の概要は準備中です。原文の説明を表示しています。

Use when designing or fixing a Content Security Policy on a real site, choosing between nonce-based and hash-based CSP, adding strict-dynamic, debugging "Refused to execute inline script" errors, deploying CSP in report-only mode first, configuring report-to / report-uri, or auditing an existing policy for unsafe-inline / unsafe-eval / wildcards. Triggers: "CSP blocks legitimate inline script", strict-dynamic, nonce-{RANDOM}, sha256-{HASH}, object-src none, base-uri none, frame-ancestors, Trusted Types, X-Content-Security-Policy obsolete, report-only vs enforced. NOT for general HTTP security headers (HSTS, COOP/COEP), Trusted Types deep dive, CORS configuration, or building a WAF.

日本語の概要は準備中です。原文の説明を表示しています。

Treat tokens as a coding swarm's cost-of-goods-sold and legibility engine while accounting for finite subscription allowance when per-call price is unknown. Covers provider windows, remaining-usage evidence, burn forecasts, preemptive checkpoint/model switching, per-agent budgets, compaction, shared digests, context degradation, and spend metering. Activate on: "token budget", "context budget", "compaction strategy", "briefing as compression", "context rot", "summarization collapse", "COGS for agents", "subscription usage remaining", "five-hour or weekly limit", "burn forecast", "model switch before limit", or "/context-economics-for-agent-swarms". NOT for: memory architecture (use always-on-agent-architecture), single-prompt wording (use prompt-engineer), mechanism-design proofs (use nisan-et-al-2007-algorithmic-game-theory), physical containment (use sandboxed-adversarial-test-harness), or production rebodiment and fencing (use agent-resurrection-and-body-continuity).

日本語の概要は準備中です。原文の説明を表示しています。

Designs and audits native-unit capacity reservation, serial admission, ambiguity holds, stop reserves, effect closure, and control-disjoint settlement handoff for consequential agent attempts. Use when scarce capacity must be conserved across retry, rework, resurrection, or compensation. NOT for scheduling work, choosing models, pricing unlike resources into one scalar, authorizing effects, grading work, paying workers, or asserting runtime containment.

日本語の概要は準備中です。原文の説明を表示しています。

Diagnose when intuitive judgment, agent confidence, or expert routing can be trusted by classifying environment validity, feedback quality, and task-boundary fit. Use for confidence calibration, agent routing, expertise audits, and escalation design. NOT for deterministic implementation tasks, pure syntax debugging, or domains with explicit verifiable answers.

日本語の概要は準備中です。原文の説明を表示しています。