5,655 件 ・ 関連度順
概要と使いどころ
Axis-by-axis interview methodology for establishing engineering performance budgets during architecture PRD compilation. Use when conducting the Step 10.5 performance budget interview in /create-prd-compile.
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。
Systematic methodology for verifying that IA, BE, and FE specification layers are mutually consistent. Covers coverage matrices, field mapping checks, error code propagation, access control consistency, and data contract drift detection.
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。
Leverage concurrent terminal background processes to debug multiple independent failures simultaneously using ACH methodology. Use when you have 2+ failures across different subsystems.
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。
Design direction interview methodology — SUPERSEDED. The `impeccable` skill's `/impeccable teach` command now handles design direction discovery with a more comprehensive, multi-round interview that writes PRODUCT.md and DESIGN.md. This file is kept for pipeline backward compatibility — all new design work should use `/impeccable teach` instead.
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。
Structured methodology for adversarial thinking — generating attack scenarios, abuse cases, race conditions, and security edge cases against specs and implementations. Produces spec-level gap items, not code-level fixes.
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。
How to upgrade `alchemy` and the `effect` constellation across this repo and keep them consistent for consumers: which packages pin what, why `effect` and every `@effect/*` companion move as one set, what breaks in a typical upgrade, and how to verify a standalone `npm install` still resolves a single `effect`. Use when asked to upgrade or bump alchemy or effect, when a deploy dies inside an alchemy provider with a `TypeError` naming a missing combinator, when `check:npm-effect-resolution` fails, or when deciding whether the alchemy patch is still needed.
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。
Enforce immediate execution of any commitment made in a reply. Use this skill before finalizing ANY reply that contains a promise to act — phrases like "אדווח", "אעדכן", "אשלח", "אחזור אליך", "I'll send", "I'll report", "I'll follow up", "I'll update". Prevents the agent from saying it will do something and then not doing it.
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。
Turn owner feedback about agent behavior into concrete system changes. Use when the owner says something is off, wants the assistant to improve how it operates, asks for a reflection, or wants a durable fix instead of a one-off apology.
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 working with Braintree — braintree payment gateway management including transactions, subscriptions, customers, payment methods, disputes, settlements, and merchant account configuration. Covers transaction success rates, decline analysis, settlement reporting, and fraud tool monitoring.
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 working with Clerk — clerk authentication and user management platform covering users, sessions, organizations, invitations, and sign-in methods. Use when analyzing user growth, monitoring authentication health, reviewing organization membership, managing Clerk users, or auditing sign-in patterns.
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 ideas via named methods from creative practice.
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。
LLM) can be manipulated through prompt injection, whether a retrieval pipeline leaks confidential da
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。
Consciousness and AI — OECS — 5. Feb. 2026 · Whether AI consciousness is possible depends on what pr GAP: OpenHands has a modular SDK design — our too
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。
A collection of methods in an intelligent software structure allows the user to focus on fine-tuning
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。
In each setting, LLM-based methods substantially outperform classical baselines, achieving up to 68% recall at 90% precision compared to near 0% for t
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。
Training Details of the Second RL Stage 4 Experiment 5 Ethics and Safety Statement 6 Conclusion, Limitation, and Future Work 7 Author List A Backgroun
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。
Autonomously improve a real artifact (code, training recipe, agent harness, data pipeline, prompt) against an objective and an evaluator, using Hypothesis Tree Refinement (HTR) from the Arbor paper. Use this whenever someone wants to iteratively optimize something over many experiments without overfitting — e.g. "get my model's eval score up", "improve this agent/harness", "tune this pipeline", "beat the baseline on this benchmark", "run a search over approaches and keep the best", "do an MLE-bench / Kaggle-style optimization", or any long-horizon "make this artifact better and don't just memorize the dev set" task. Trigger it even when the user doesn't say "Arbor" or "hypothesis tree" but describes repeated experiment-and-evaluate loops, branching exploration of competing ideas, or worries about a dev/test gap. Runs Claude itself as the coordinator with subagent executors in isolated git worktrees; for the standalone `arbor` CLI tool see references/arbor-upstream.md.
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。
Multi-perspective academic paper review with dynamic reviewer personas. Runs a 5-seat, role-separated review panel (Journal-Fit Reviewer + 3 peer-review roles + Devil's Advocate) with field-specific expertise; role separation is not a claim of independent error processes. Supports full review, re-review (verification), quick assessment, methodology focus, Socratic guided, and calibration modes. Triggers on: review paper, peer review, manuscript review, referee report, review my paper, critique paper, simulate review, editorial review, calibrate reviewer, reviewer calibration, measure reviewer accuracy, 審查論文, 論文審查, 模擬審查, 同儕審查, 幫我審這篇, 以審查人角度評估, 審查者校準, 논문 심사, 동료 심사, 모의 심사, 심사자 관점에서 평가, 심사자 보정.
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。
Stop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.
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。
Never test mock behavior. Never add test-only methods to production classes. Understand dependencies before mocking.
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。
Java Javadoc best practices. Use when adding or reviewing documentation for Java types, methods, packages, and public 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 a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.
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。
Train or fine-tune language and vision models using TRL (Transformer Reinforcement Learning) or Unsloth with Hugging Face Jobs infrastructure. Covers SFT, DPO, GRPO and reward modeling training methods, plus GGUF conversion for local deployment. Includes guidance on the TRL Jobs package, UV scripts with PEP 723 format, dataset preparation and validation, hardware selection, cost estimation, Trackio monitoring, Hub authentication, model selection/leaderboards and model persistence. Use for tasks involving cloud GPU training, GGUF conversion, or when users mention training on Hugging Face Jobs without local GPU setup.
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 removes infostealer malware (credential stealers, data exfiltrators) via full-system file search, cryptographic hashing, and public threat-intelligence cross-checks (VirusTotal, MalwareBazaar). Primary method is always custom hash-based detection. Windows Defender (or any platform-native AV) is allowed **only when necessary** (e.g. inconclusive hashes or deep remediation) and **must never be the default option**. The agent must exhaust the custom workflow first. Works on Windows/macOS/Linux.
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。