3,758 件 ・ 関連度順
概要と使いどころ
Apple Silicon GPU コンピュート／数値計算リファレンス。MSL の kernel 関数属性・address space・ atomic/SIMD-group 関数、threadgroup sizing・argument buffer・indirect command buffer・ unified memory・GPU counters、MPS/MPSGraph（MPSMatrix, MPSNDArray, MPSNNGraph）、 MLX（mx.array, lazy evaluation, grad/vmap, mx.compile, mlx.nn, mlx.optimizers, distributed）。
Swift Standard Library・Concurrency・プロトコル/ジェネリクス・Observation リファレンス。 async/await、actor、Sendable、Task、TaskGroup、AsyncSequence、AsyncStream、 Equatable、Hashable、Comparable、Codable、Identifiable、Optional、Result、 generics、some/any opaque types、Observable、@Observable、ObservationRegistrar。
Debrief a finished Otameshi Tenshoku trial from the employer's side: set the observed facts against the completion criteria agreed before the start, keep expectations added afterwards, second-hand or off-the-job observations and the company's own late or missing support apart, settle the hours actually worked separately from any evaluation, and draft feedback to the candidate that states facts, support gaps and what happens next without promising a hire or continuation the company has not decided. Use when a hiring manager, recruiter or small-business owner wants to review a trial, decide what to tell the candidate, or work out the settlement after a shortfall; do not use to judge the candidate's aptitude or predict a hire, to cut pay or demand unpaid rework because of dissatisfaction, to compare candidates, or to send the feedback, pay, or update records on the user's behalf.
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。
Claude Code のセッション記録（~/.claude の履歴・セッションログ・memory）から、週ごとの問い（questions.json）に答える「ナレッジカード」を下書きし、本人が確認・修正したうえで機密スキャンを通し、Notion の「ナレッジカード」DB に提出する。引数なしなら先週（月曜〜日曜）の記録を集計し、確認のうえ Notion に提出する。「Claude Code の履歴をナレッジカードにまとめて」「先週のログからナレッジカードを作って」「ナレッジカードを提出したい」「今週のカードを出して」「週次の振り返りカードを作って」などの依頼で使う。週次のカード提出であり、個人の採点や評価ではない（セッション 1 本の採点は evaluate-session の領域）。
週ごとに Notion「ナレッジカード」DB に集まったカードを読み、問いをまたいで親和図にまとめ、正（worked）と反（failed）の対立から合（提案）を出し、次の問いを決めて、「合成テーマ」DB と週の行の配下の「まとめ」ページに書き込む。「今週のカードを合成して」「W43 のまとめを作って」「ナレッジカードの合成」などの依頼で使う。カードの提出は knowledge-cards、個人の採点は evaluate-session の領域。個人別の件数ランキングや「誰が何を失敗した」は出さない。
Quality control for pooled CRISPR screens covering library representation, Gini index, log-skew, replicate Pearson and Spearman concordance, essentialome precision-recall AUC against CEGv2 (Hart 2017), Cas9 cut-toxicity diagnostics, copy-number amplicon detection (Aguirre 2016 / Munoz 2016), bottleneck propagation through plasmid pool, infection, selection, and endpoint stages, MOI verification, and DepMap-style screen-quality scoring. Use when assessing screen quality before hit calling, deciding whether to repeat or rescue a screen, diagnosing low-confidence hits, choosing between MAGeCK / BAGEL2 / Chronos based on quality grade, picking a normalization strategy from QC signatures, or evaluating whether an in-vivo screen retained adequate library complexity.
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。
Track and report progress of a request to create and run an EcoSIM case for a site ("create an ecosim run for site US-UMB", "set up and run EcoSIM at <site>"). Records four stages — data preparation, namelist creation, model simulation, and output evaluation — in result/<SITE_ID>/progress.json and PROGRESS.md, derives step status from files and quality reports on disk, and records manual results such as the executable, test run, and evaluation verdicts. Use at the start of such a request, after each stage, and whenever the user asks where a site run stands.
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。
Provides patterns to configure Spring Boot Actuator for production-grade monitoring, health probes, secured management endpoints, and Micrometer metrics across JVM services. Use when setting up monitoring, health checks, or metrics for Spring Boot 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。
Evaluate design from a UX perspective, assessing visual hierarchy, information architecture, emotional resonance, cognitive load, and overall quality with quantitative scoring, persona-based testing, and actionable feedback. Use when the user asks to review, critique, evaluate, or give feedback on a design or component.
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。
6 production-ready AI engineering workflows: prompt evaluation (8-dimension scoring), context budget planning, RAG pipeline design, agent security audit (65-point checklist), eval harness building, and product sense coaching.
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。
スキル使用後の observation（結果・問題・フィードバック）を記録する。 スキルの改善サイクル（Observe → Inspect → Amend → Evaluate）の起点。 使用後に `/skill-observe <スキル名> <結果> [問題/フィードバック]` で呼び出す。
蓄積された observations を分析し、SKILL.md への改善提案を生成する。 Observe → Inspect → Amend → Evaluate ループの Inspect/Amend フェーズ。 `--apply` フラグで承認後の適用も可能。
Claude Code の過去のセッション記録（~/.claude の履歴・セッションログ・memory）から、他のメンバーと共有できる運用 Tips を抽出し、公式ドキュメントや論文などの一次情報で裏付けを取って、Markdown・Artifact・Notion に公開する。「過去の会話から Tips をまとめて」「Claude Code の使い方をチームに共有したい」「今月の学びを Tips にして」などの依頼で使う。セッション 1 本の効率採点ではない（それは evaluate-session の領域）。週次のナレッジカード提出は knowledge-cards で行い、このスキルはカードや記録から Tips 文書を作る下流工程にあたる。
Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, 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。
Build, test, and debug Hermes Agent RL environments for Atropos training. Covers the HermesAgentBaseEnv interface, reward functions, agent loop integration, evaluation with tools, wandb logging, and the three CLI modes (serve/process/evaluate). Use when creating, reviewing, or fixing RL environments in the hermes-agent repo.
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。
Generate 5–10 business idea candidates from a blank page or a founder's domain context — using pain mining, jobs-to-be-done, trend × capability mapping, constraint relaxation, adjacency search, and founder-market-fit prompts. Each candidate is a structured idea card (segment, JTBD, current alternative, why-now, distribution wedge, monetisation, "feels like"). Load when the user asks to generate business ideas, brainstorm startup ideas, find ideas to work on, says "what business should I start", "give me startup ideas", "I don't know what to build", "ideate ventures", "blank-page idea generation", "find me a startup idea", "explore business opportunities". Sub-skill of `venture-exploration`. Hard-bans "Uber for X" / "AI for X" with no specific JTBD, "everyone" segments, and idea cards missing any of the 7 required fields. Does NOT design or evaluate ideas generated — for that use `idea-evaluation`.
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。
Internal skill. Called by setup-evaluation after a PASS. Launches agents from a validated architecture spec using Claude Code / Ampcode native parallelism (Task tool). Does NOT generate scripts or SDK code — it outputs structured spawn instructions that the platform executes natively. Never invoked directly by the user. Never launches without a setup-evaluation PASS.
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。
6 production-ready AI engineering workflows: prompt evaluation (8-dimension scoring), context budget planning, RAG pipeline design, agent security audit (65-point checklist), eval harness building, and product sense coaching.
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。
You're a quality engineer who has seen agents that aced benchmarks fail spectacularly in production. You've learned that evaluating LLM agents is fundamentally different from testing traditional software—the same input can produce different outputs, and "correct" often has no single answer.
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。
Quality control for pooled CRISPR screens covering library representation, Gini index, log-skew, replicate Pearson and Spearman concordance, essentialome precision-recall AUC against CEGv2 (Hart 2017), Cas9 cut-toxicity diagnostics, copy-number amplicon detection (Aguirre 2016 / Munoz 2016), bottleneck propagation through plasmid pool, infection, selection, and endpoint stages, MOI verification, and DepMap-style screen-quality scoring. Use when assessing screen quality before hit calling, deciding whether to repeat or rescue a screen, diagnosing low-confidence hits, choosing between MAGeCK / BAGEL2 / Chronos based on quality grade, picking a normalization strategy from QC signatures, or evaluating whether an in-vivo screen retained adequate library complexity.
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。
Apply recognition-primed decision making to agents operating under time pressure, uncertainty, and ambiguous cues. Use for situation assessment, RPD-style action selection, and decision-support design in operational settings. NOT for batch optimization, static planning with full information, or tasks where exhaustive search is cheap and harmless.
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。
Apply recognition-primed decision making to agents operating under time pressure, uncertainty, and ambiguous cues. Use for situation assessment, RPD-style action selection, and decision-support design in operational settings. NOT for batch optimization, static planning with full information, or tasks where exhaustive search is cheap and harmless.
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 deterministic, authority-filtered partitions of trust-typed context for already admitted bodies or abstract continuation slots. Use when causal context, obligations, disclosure boundaries, vector-space identity, capacity, and omission proofs must survive partitioning. NOT for spawning or admitting agents, choosing worker count, retrieving arbitrary knowledge, writing successor prompts, or granting tools, leases, identity, or effect authority.
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。