「grading」の検索結果

Assesses ChIP-seq quality across antibody specificity, fragmentation, enrichment, replicate concordance, and library complexity. Computes FRiP, NSC/RSC (phantompeakqualtools), library complexity (NRF/PBC1/PBC2), deepTools plotFingerprint (JS distance, AUC, synthetic JS), ChIPQC, IDR with ENCODE Nself/Nt rules, and detects hyper-ChIPable artifacts. Use when validating an antibody, diagnosing failed peak calls, deciding whether to proceed with downstream analysis, grading against ENCODE thresholds, or auditing replicate concordance.

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 upgrading Stripe API versions, webhook endpoints, server-side SDKs, Stripe.js, and mobile SDKs

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。

Writes turn-level tests for a LiveKit agent in the user's normal test suite: pytest (Python) or Vitest (Node.js). Use when the user asks to "write tests for my agent", "add a test for this tool", "test the handoff", "pin this bug", "why does my agent test fail", or after building or changing agent behavior that needs regression coverage. Covers the SDK's test session harness, assertions on messages, tool calls and handoffs, LLM judging of a reply against an intent, mocking tools, multi-turn tests, and judging whole conversations with the built-in judges. For interactive poking use debugging-livekit-agents. For grading whole conversations at scale use running-livekit-simulations.

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 moving skills between library workspaces or upgrading from a personal library to a team library. Export from one workspace, import into another.

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。

Detect quality and efficiency regressions over time using Agent Monitor data — rising error rate (APIError events), falling cache hit rate, growing compaction frequency, and climbing cost-per-session. Splits history into an earlier baseline window and a recent window and reports which metrics are getting worse, by how much, and where. Use when checking whether things are degrading or trending in the wrong direction.

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 improve an entire agentic-development setup, including skills, context, memory, tools, models, permissions, hooks, workflow habits, and secrets hygiene. Invoke for setup reviews, workflow grading, Goblin Mode, or questions about what to improve. Supports shallow, deep, ultra, and yolo modes. Use domain-specific skills for a single model, report, or tenant.

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 making multiple independent ad or campaign edits from existing footage with Scenario: hook, CTA, copy, end-card, placement, or targeted visual variants. Also for revising one variant or resuming a partially completed batch. Not for a new commercial from a still image, ordinary single-clip grading, or automatic Cartesian combinations.

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 lighting or rendering in Unreal Engine 5.8: physical light units and exposure (EV100, lux, lumens), Lumen GI and reflections, MegaLights, virtual shadow maps, sky, fog and clouds, post process and color grading, the path tracer, or Movie Render Graph stills. Triggers: "Lumen is noisy", "MegaLights noise", "the interior is too dark", "light leaks", neon signs, a night or dusk scene, "fireflies", "path tracer vs Lumen", "render a still", EXR settings.

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 setting up URP (or choosing URP vs HDRP), configuring URP assets and quality tiers for PC and phones, lighting a stylized or realistic Unity scene (sun, sky, baked or mixed lightmaps, Shadowmask, Adaptive Probe Volumes, reflection probes), or fixing blurry shadows, light leaks, gray bakes, washed-out images, tonemapping and grading, bloom too expensive on mobile, TAA, STP or MSAA choices, GPU Resident Drawer setup, shader warm-up hitches, or a custom shader that vanished or lost its baked light.

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 editing footage that already exists on Scenario through MCP with a tool model instead of a new generation: 3D LUT grading, color correction, film grain, vignette, blur, sharpen, glow, tint, desaturate, posterize, solarize, cubism, crystallize, oilify, chromatic aberration or dodge and burn on a clip; trimming to a time range, splitting at cut points, reversing, resizing, re-encoding, frame extraction, image sequences, background removal, segmentation masks, subjects as layers.

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。

Cluster API lifecycle management for provisioning, scaling, and upgrading Kubernetes clusters. Use when managing cluster infrastructure or multi-cluster operations.

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。

Analyze images, video, speech, motion, products, and websites; route local vision, audio, depth, tracking, matting, identity, and restoration models; auto-edit, replicate, enhance, caption, voice, assemble, annotate, validate, and export editable Timeline Studio projects and videos. Use for reference-video remakes, filter and repeated-shot reconstruction, subject-aware reframing, clip splitting, source-time speed curves, Color Wheels grading, ramps and holds, person or product cutout, person or object outline, authorized face swap, shot and timing reconstruction, timeline markers, chapters, beat cues, review notes and ranges, supplied or web-sourced footage, AI video platform selection, cleanup, highlights, product promotion, website walkthroughs, image-to-video assisted edits, optical-flow editing, depth/2.5D/transition finishing, AI voiceover, subtitles, short-form production, .timeline automation, or editor 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。

Live-site visual QA audit with fix loop. 10-category checklist (~80 items) including AI Slop Detection (10 blacklisted patterns). Atomic commits per fix. Before/after screenshots. A-F grading. Parallel cross-model code-level design audit. Modes: Full, Quick, Deep, Diff-aware, Regression.

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 custom LLM evaluation pipelines using the OpenJudge framework. Covers selecting and configuring graders (LLM-based, function-based, agentic), running batch evaluations with GradingRunner, combining scores with aggregators, applying evaluation strategies (voting, average), auto-generating graders from data, and analyzing results (pairwise win rates, statistics, validation metrics). Use when the user wants to evaluate LLM outputs, compare multiple models, design scoring criteria, or build an automated evaluation system.

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。

Verifies UK case-law citations, pinpoint references and quotations against the official public register — The National Archives' Find Case Law — and statutory references against legislation.gov.uk, before a document that cites them is relied on, served or filed. Every check returns a graded verdict: VERIFIED, MISMATCH (the citation resolves to a different case — the classic AI miscitation), NOT ON REGISTER, OUTSIDE COVERAGE (the register cannot answer — absence proves nothing), or UNCHECKABLE (a law-report citation). The grading exists because there are two ways to get this wrong: citing a case that does not exist, and accusing a real case of not existing. Use when asked to check citations, verify a case exists, confirm a quotation is verbatim, check a pinpoint paragraph, audit a draft's authorities, screen a document for hallucinated cases, or check a statutory provision is in force. Not a substitute for reading the judgment: existence is not 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。

Turns a legal question, a set of draft questions, or source material such as a statute, jury instruction, rubric, contract checklist, or coding manual into a battery of narrow questions that TypeSafe's Jev, a fast classifier that returns probabilities rather than prose, can answer reliably across many documents. The skill breaks conclusions such as negligence, felony murder, or standing into element-level questions, writes the governing rule into each question, and delivers the questions alone, in Jev's request format or in plain English. Built for legal work that asks the same questions of many texts: coding statutes across jurisdictions, coding opinions or case files for empirical research, checking contracts against a review checklist, screening intake narratives, and grading student work against a rubric. Use when asked to write Jev questions, build a classifier for legal documents, or turn a legal test into yes-or-no questions. Running the questions requires Jev access; writing them does not.

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 *before* starting work in a domain you don't know well, to surface the "unknown unknowns" — the things you don't even know to ask about — and learn just enough to prompt and decide well. Implements the "blind spot pass" pattern from Anthropic's Fable "finding your unknowns" field guide. Triggers when you say "I'm new to this", "I don't know what to ask", "teach me before we start", "blind spot pass", "unknown unknowns", "find my unknowns", or the Korean "내가 뭘 모르는지 알려줘" / "이 분야 처음인데" / "먼저 가르쳐줘" / "블라인드 스팟", or when you hand off a task while admitting you're a non-expert in that field (color grading, video editing, legal, tax, finance, design, an unfamiliar codebase, etc.). Do NOT use for simple factual questions, domains you already know, or directly-actionable work — answer those directly or route to planner / architect.

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。

Manage major dependency version upgrades with compatibility analysis, staged rollout, and comprehensive testing. Use when upgrading framework versions, updating major dependencies, or managing breaking changes in libraries.

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 a WRONG or imperfect render (not a hard error) by inspecting inputs and intermediate steps with run-to-node. Render one branch up to an output, preview-tap latents/masks/preprocessor maps, localize the first bad stage, then fix. Use when a final image/video completes but looks wrong, such as artifacts, wrong subject/pose/composition/color, blur, a ControlNet/IPAdapter/mask/LoRA not taking, or a two-stage refiner or upscale degrading the result. For errors/OOM/missing nodes use the troubleshooting skill.

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。

Detect and interrupt the pattern where an agent confidently praises work it just produced instead of reviewing it critically. Same-context grading is not review — it's rationalization.

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 upgrading Java projects between major versions (8→11→17→21→25). Use when user says "upgrade Java", "migrate to Java 25", "update Java version", or when modernizing legacy projects.

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。

Analyze skill effectiveness across sessions. Computes per-skill metrics (action rate, friction, outcomes), identifies degrading skills, and generates improvement recommendations. Requires session-scan data in metrics.jsonl.

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。

Helps a teacher prepare to teach an existing K-12 lesson (math / ELA / science / social studies) — a prep partner that thinks through the lesson's key student task with the teacher and leaves a short teacher-only prep note. Triggers on asks to internalize, prepare for, get ready to teach, or understand a lesson, including implicit ones ("help me get ready for tomorrow's lesson"), and right after another skill built a lesson, which it can follow. Not for building or adapting a lesson, grading, rubrics, assessment feedback, or generating quizzes or student-facing material, and not for single-fact lookups a sentence would answer ("what standard is this?", "how long is this lesson?").

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。

Adapts an existing K-12 lesson (math, ELA, science, or social studies) for students at different proficiency levels (below / at / above grade level). Load this skill BEFORE asking the teacher any clarifying question about the lesson, tiers, or student levels. Triggers on explicit asks to differentiate, tier, or scaffold a lesson, and on implicit signals like "my students are at different levels". Produces 1 teacher-facing differentiation plan + 3 student-ready tier documents as editable Word documents in a single output turn, rendered from one material-source JSON via bundled scripts (shared content is written once so tiers cannot drift). Uses the Learning Commons Knowledge Graph when connected; works without it. This skill adapts a lesson the teacher brings or names. Not for creating a new lesson from scratch — a new-lesson request that asks for differentiated or leveled materials is k12-lesson-plan-creation's job, one package. Not for grading, rubrics, assessment feedback, or quizzes.

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。