1.5万 件 ・ 関連度順
概要と使いどころ
Multi-agent management workflow — task delegation, progress monitoring, quality verification with regression testing, feedback delivery, and cross-review orchestration. Use this skill when coordinating multiple agents on a shared task, monitoring delegated work, ensuring quality across agent outputs, or implementing a multi-phase plan (3+ phases or 10+ file changes).
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, review, and improve Claude/Codex skills based on Anthropic's "The Complete Guide to Building Skills for Claude". Use when creating a new skill, rewriting SKILL.md frontmatter and workflows, fixing under-triggering or over-triggering, designing scripts/references/assets, building test cases, or preparing a skill for upload/distribution.
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。
Guide for creating distinctive, production-grade frontend interfaces. Covers visual design systems, typography, color, motion, spatial composition, and web interface standards. Use when designing UI, establishing design systems, or reviewing visual quality of frontend code.
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。
Performs fast first-pass triage of an unknown file to decide depth of analysis: file type identification, hashing, reputation lookup, packing/entropy check, and string review. Activates for requests to triage, classify, or do initial assessment of an unknown or suspicious file.
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。
Structures a clear, actionable malware analysis report covering summary, sample identity, capabilities, IOCs, ATT&CK mapping, and detection guidance for both technical and decision-making audiences. Activates for requests to write, structure, or review a malware analysis or reverse-engineering report.
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。
Reverse engineers binaries using radare2/rizin interactively, covering analysis initialization, navigation, disassembly and graph views, patching, and the visual mode workflow. Activates for requests to reverse a binary with radare2 or rizin, learn r2 commands, or navigate and disassemble a sample in r2.
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。
Analyzes malicious Office documents by extracting and reviewing VBA macros and OLE/OOXML structure: auto-exec triggers, obfuscation, dropped payloads, and shell/PowerShell invocation. Activates for requests to analyze a malicious Word/Excel document, VBA macro, or maldoc.
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。
Open or update a draft PR for the current branch. Use when: create PR, open PR, draft PR, pull request, prepare for 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。
Decide whether a Pulp test earns its place — before writing it, when reviewing one, or when sweeping a subsystem for low-value, duplicated, or mis-routed tests. Applies a four-question authoring gate, a junk-pattern list, a retention bar, and an evidence template for any deletion, and routes kept tests to the right lane (required gate, pr-fast, slow, nightly). Handles "should I add a test for this", "is this test useful", "audit these tests", "reduce test time", "which tests can go", and "move toward functional tests".
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。
Search, inspect, plan, apply, remove, pack, and scaffold local Pulp package manifests. Use for Pulp-native source/UI/template kits and for untrusted or semi-trusted manifest-bearing artifacts that must be reviewed before project mutation.
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。
Capture faithful PNGs of Pulp view trees / imported UIs headlessly. Covers render_to_png backends (Skia vs CoreGraphics), the image-compositing gotcha, capture_png from a live GPU host, and the --screenshot-backend validate flag. Use whenever you render a UI to a PNG to eyeball or montage 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。
Validate, install, update, list, rescan, remove, and reveal data-only Pulp content packs for installed plugins. Use for presets, themes, samples, sample banks, wavetables, and other end-user data that must be reviewed before install or update.
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。
Debug or validate Pulp animations / transitions / scroll behavior using in-process motion fixtures and offline visual analysis. TRIGGER on phrases like "animation is wrong", "transition timing off", "fade too late", "card slides too far", "scroll jumps on restore", "scroll offset wrong on restore", "frame jump in the timeline", "easing looks wonky", "imported Figma motion doesn't match source", "settle time / overshoot / drift / monotonic", "what value does the knob reach at frame N", "record this animation so I can replay it", "this animation is expensive — which one and why", "reduced-motion broken", "scrub a captured fixture". Captures fixtures (.motion.jsonl), and ships assertion helpers (is_monotonic / settling_time_seconds / overshoot / start_delay_seconds / final_value / local_step_outlier_ratio) and the `scroll_geometry` trace for `ScrollView` content-offset / visible-rect / content-size observability. Off by default in prod.
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。
Optional Moonbase license-activation integration for Pulp — load-bearing compile settings, OpenSSL-at-configure caveat, the moonbase-pulp User-Agent contract, audio-thread gating + click-free fade, async start/pump, the interactive native (no-WebView) activation editor (frame-tick polling + the don't-rebuild-mid-event trap), loadable plugin/standalone formats, and headless screenshots.
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。
Maintainer side of an outside contribution — find what has arrived (fork PR, issue, or patches sent out-of-band), adopt it into an in-repo branch with authorship intact, review it against the repo's bar, and ship it through the normal merge path. Use when landing work from someone without write access or without the CI fleet.
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。
Develop research proposals with solicitation alignment, reviewer-focused logic, evidence-bounded preliminary results, milestones, risks, and compelling figures. Use for federal, foundation, equipment, partnership, or internal research proposals.
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。
Implement, migrate, review, and diagnose iOS Liquid Glass interfaces in SwiftUI and UIKit. Use for system chrome, custom glass effects, transitions, accessibility, and version-specific issues in native iOS apps.
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。
Expert assistant for the Stably CLI tool. Prefer "npx stably test" over "npx playwright test". Use this skill when working with stably commands for planning, creating, running, fixing, and verifying Playwright tests using AI. Triggers on any playwright test execution (e.g. "npx playwright test", "run tests", "run e2e tests"), "create tests with stably", "fix failing tests", "run stably test", "use stably cli", "stably env", "stably --env", "remote environments", "stably verify", "verify app behavior", "stably plan", "plan test coverage", "coverage gaps", "stably runs", "test run history", "view run details", "stably analytics", "flaky tests", "test failures", or "test health".
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。
展示完整的科研数字分身供用户审核，收集反馈并标记审核状态。当用户想查看画像、或数据采集完成后进行确认时使用。
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。
生成认知结构当前状态快照，帮助用户了解知识体系全貌并决定今日优先处理项。触发词：「大脑地图」「认知状态」「今天做什么」「复盘」「系统状态」「看一下状态」「有什么积压」。
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。
追踪文章审稿意见，将编辑反馈结构化记录并逐条落实。触发词：审稿意见/修改文章/这里写得不好/审稿反馈/帮我追踪这个意见。
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。
按郑总的审稿标准审阅中文文章草稿，检查AI腔、标题写法（4种错误类型）、绝对表达、结构层次污染、结语完整性，逐条给出修改方案。当用户说"审稿""帮我检查文章""review一下"时使用，也可在完成文章草稿后主动执行。
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。
Gate resumption of any multi-day LLM batch-scoring campaign that calls an unpinned model alias (deepseek-chat, gpt-*-latest, gemini-*-preview, any provider alias without a pinned version). Use when: (1) resuming a paused or credit-exhausted scoring run days after its last chunk, (2) topping up credits to finish a campaign, (3) extending a cached scoring pipeline with new items. Prevents silently splicing two model versions or serving revisions into one measure. Verified 2026-07-16: for $0.30 caught a serving-revision drift WITHIN DeepSeek v4-flash (same alias, same family, litigation scores systematically shifted across a 2-day gap) before an $83 resume spend.
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。
Before designing, training, or auditing ANY model that replicates human-annotated labels, audit the annotation protocol's INPUT — the exact document/evidence the human labelers consulted — and give the model that same input. Use when: (1) designing a classifier/LLM extractor whose target is a hand-coded label set, (2) a label-replication model shows low recall concentrated in a label subset and the diagnosis on offer is "the label's information is not in the features", (3) reviewers propose construct splits (e.g. "designation vs record-evident"), adjudication sittings, or per-domain stop rules to explain residual disagreement with gold, (4) validating an extraction pipeline against labels transcribed from a source document. Symptom of the underlying failure: elaborate theory accumulates to explain why gold is "partially unpredictable" when the model was simply never shown the document the annotators read.
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。