Claude Skills
The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 79 566 files from 1 758 authors, of which 61 913 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Model family compatibility matrix — loaders, resolutions, samplers, CFG, VAE, ControlNet, and LoRA compatibility for SD 1.5, SDXL, Flux, SD3, and video models
Drive the LTX Director (Timeline) node — its Add Image/Text/Audio buttons are DOM-only and cannot be clicked by an agent; edit the hidden timeline_data JSON widget instead. Load when a workflow contains LTXDirector / LTXDirectorGuide / PromptRelayEncodeTimeline, or when asked to add, move, retime, or remove timeline segments (image / text / audio / motion).
ComfyUI prompt engineering knowledge — CLIP text encoding syntax, weight modifiers, model-specific prompting strategies, and best practices
Build Qwen Image Edit workflows — model loading, conditioning, LoRAs, prompt patterns, and XY plot testing
Keep the ComfyUI sidebar panel node-pack (comfyui-agent-panel) in step with the orchestrator after comfyui-mcp updates. Use this whenever the orchestrator was just updated (self_update, npm i -g comfyui-mcp, a new version in the ENVIRONMENT line), when a panel/bridge command fails in a way that smells like version drift ("panel is too old", a graph_/ui_ command the panel doesn't implement, a feature that works in the docs but not in the sidebar), or when the user asks to update/pin/unpin the panel. It checks the installed panel version against what THIS orchestrator build needs, RESPECTS an explicit version pin (warn-only, never move a pinned user), offers a clear way to unset the pin, runs the sync through the verified install_panel path, and reports the version RE-READ from disk. Never claim a sync that did not happen.
Self-heal + report bugs to make the ecosystem better. When you hit ANY defect in comfyui-mcp / the sidebar panel, OR in a third-party custom node / ComfyUI core, use this: diagnose it, and for OUR repos (comfyui-mcp / comfyui-mcp-panel) **fix it locally FIRST** — patch the code where it actually runs so the user is unblocked immediately — then file a report that includes the diff, so reports arrive as near-PRs, not just tickets. (Attempt the fix exactly ONCE; if it's genuinely upstream-only, mark it and include the precise change needed.) Report it the right way: our intake Worker for our repos; the node's own GitHub for third-party. For OUR repos (artokun/comfyui-mcp, comfyui-mcp-panel, comfyui-mcp-issue-worker) act **autonomously** — fix-then-file and just inform the user, no per-step permission. For THIRD-PARTY / ComfyUI-core it is **offer-and-ask**: propose the workaround/report and act only with the user's go-ahead. **BETA — bias HARD toward fix-then-file (our repos):** trigger on any comfyui-mcp/panel defect you notice — an error, a crash, a tool that returns wrong/misleading/empty output, a silent failure, a confusing result, or a limitation that forced an awkward workaround — EVEN IF you recovered or worked around it. You do NOT need to be blocked. Also on "report this" / "fix this bug". If a future user would hit the same friction, fix-and-file it — the intake Worker dedupes, so over-reporting is cheap and under-reporting costs us the bug. THIRD-PARTY / ComfyUI-core bugs stay offer-and-ask.
Build Qwen Image 2512 text-to-image workflows — QwenImageIntegratedKSampler, separate component loading, lightning LoRAs, and fine-tuned model variants
Train a character/identity LoRA locally on FLUX.1-dev via the comfyui-mcp train_* tools (GPU Docker + ostris ai-toolkit). Use when the user wants to train a LoRA of a person/character from their photos on the local GPU — covers dataset prep, launch, monitoring, and using the result in ComfyUI. For WAN/Z-Image training via the ai-toolkit UI see ai-toolkit-trainer.
Upscale and restore video in ComfyUI — both the quick local path (per-frame ESRGAN like 4x_foolhardy_Remacri via ImageUpscaleWithModel + 4x→2x supersample, with its temporal-flicker tradeoff) and temporal-aware super-resolution (SeedVR2, the newer FlashVSR) with the downscale-first restore pipeline; RIFE/FILM frame interpolation via the BUILT-IN ComfyUI 0.26 FrameInterpolate (rife_v4.26 in models/frame_interpolation/) or the ComfyUI-Frame-Interpolation pack; 2x/4x scaling, VRAM tiers, VHS encode. Captures the classic downscale→SeedVR2→RIFE recipe and the current 2026 recommendation.
Common ComfyUI errors and fixes — OOM, missing nodes, dtype mismatches, black images, and debugging strategies
Extend / continue a video temporally with Pusa 2.2 in ComfyUI — temporal flowmatching (the flowmatch_pusa scheduler + WanVideoAddPusaNoise) on the WanVideoWrapper stack with WAN 2.2 T2V A14B (HIGH/LOW) models and the Pusa V1 LoRAs, conditioning on the loaded clip via WanVideoEncode so the existing motion carries into the continuation. Covers the kijai wanvideo_2_2_14B_Pusa_extension graph, model/LoRA slots + downloads, noise/length/scheduler settings, chaining multiple extensions, VRAM tiers, gotchas, and the extend→upscale handoff.
Install Triton + SageAttention to accelerate ComfyUI (the sageattn attention_mode and inductor torch.compile used by WanVideoWrapper / many video graphs) — Windows-first (triton-windows + woct0rdho prebuilt SageAttention wheels matched to torch/CUDA/python into the RIGHT python), plus Linux (official triton + build) and Mac (N/A → sdpa/MPS). Critically also covers the SAFE sdpa / no-compile fallback so an example that assumes sageattn + torch.compile still runs when these aren't installed (video-extend TRAP 5). Use when a loader crashes with "No module named 'sageattention'" / "triton: unavailable", when asked to speed up Wan/video workflows, or when deciding whether to install acceleration vs. fall back.
Build WAN 2.2 First-Last-Frame video workflows — native dual hi-lo (required), and WanVideoWrapper VACE approaches
SCAIL-2 in-video character replacement on WAN 2.1 — WanSCAILToVideo + SCAIL2ColoredMask + SAM3, the reference-image framing→scale rule, and the tuning/compositing pitfalls
Build WAN MultiTalk audio-driven talking-avatar / lip-sync video workflows — MeiGen-AI MultiTalk on WAN 2.1 14B I2V via kijai WanVideoWrapper (portrait + audio → lip-synced video)
Build WAN 2.2 Text-to-Video workflows — dual hi-lo models, lightning LoRAs, VACE modules, and KSamplerAdvanced two-pass
Lay out and organize a ComfyUI workflow cleanly on the live panel canvas — dependency-layered node placement with no overlaps, subgraphs, colored group boxes, and subgraph rail alignment. Use when asked to tidy / clean up / organize / arrange a workflow, add groups or subgraphs, fix overlapping nodes, or build a workflow that should look good from the start.
Build Z-Image txt2img workflows — RedCraft checkpoint, Z-Image Turbo/Base LoRAs, ControlNet, and sampler presets
Multi-engine web search (SearXNG) + browsing/scraping (Camofox, CloakBrowser). Use whenever you need to do web research.
> a service company, a physical product, a food brand, a drink brand. Scroll becomes the timeline. Video scrubs frame by frame under the wheel, sections pin and advance, rails pan sideways, headlines assemble line by line, the page ground shifts colour as you travel, and the pointer moves things that are not scrolling. Interviews the human first (their vibe, their journey, one unbroken world or distinct scenes, and what assets they already own), then picks a page grammar and a signature move so no two builds share a skeleton, generates photoreal assets through kie.ai or builds from the user's own footage and photos, writes real semantic HTML on a design-system floor, and verifies the result by screenshotting its own scroll. Use for "scrollytelling", "scroll animation site", "a site where scrolling plays a video", "Apple-style landing page", "3D scroll world", "interactive landing page", "make my brand a scroll experience", "make it feel different", "this looks like a template", "a unique scroll site", or any request for a site that should feel like an experience rather than a document.
Use when the user wants to design a mobile app or UI screens, when they mention their Sleek (sleek.design) projects, or when implementing Sleek designs in code (HTML, React Native, SwiftUI).
Design KPI frameworks, metric definitions, targets, guardrails, and measurement plans for product or business decisions. Use when success metrics, drivers, guardrails, targets, or the measurement approach need to be defined or improved.
Narrow conversion skill. Invoke only when the user explicitly asks to convert an existing Data Analytics report, dashboard, or inline chart export into a PDF artifact.
Build source-backed dashboards for monitoring performance, exploring drivers, or acting on product and business metrics. Use when the task needs a dashboard, scorecard, or monitoring view with clear metrics, filters, source definitions, and QA.
Assess whether structured data, query results, dashboards, or analytical evidence are trustworthy enough to use. Use when the task is to check data quality, reconcile conflicting sources or metric definitions, or decide whether evidence is safe to cite.
Narrow conversion skill. Invoke only when the user explicitly asks to convert an existing local or blob-hosted HTML analytics report into a Google Doc, DOCX, or shareable document.
Narrow conversion skill. Invoke only when the user explicitly asks to convert an existing HTML analytics report into a native Google Slides deck.
Create, update, inspect, or repair Data Analytics semantic layers. Use when the user asks to save data context or create a semantic layer that future Data Analytics work can inspect and cite.
Build polished analytical reports for executive, product, business, or technical audiences. Use when the task needs a durable answer-first narrative with evidence-backed findings, visuals or tables, caveats, and source context.
Gather business context from connected or provided sources so downstream analysis starts with the right framing. Use when an analytical question depends on missing context, such as what a metric means, what changed recently, or which sources should be checked. If the same prompt asks for diagnosis, recommendation, or a deliverable, gather context first and continue to the focused skill.
Route Data Analytics plugin-level requests and broad analytics work to the right focused workflow. Use when Data Analytics is at-mentioned, or for analytics requests involving data, metrics, dashboards, reports, charts, notebooks, spreadsheets, KPIs, market sizing, or semantic layers.
Diagnose why a metric changed or differs from expectation. Use when the task is to identify likely drivers of a metric movement, anomaly, gap, or discrepancy.
Estimate market, segment, or opportunity size with transparent assumptions and uncertainty. Use for TAM/SAM/SOM, sizing scenarios, or comparing the scale of possible opportunities.
Validate whether an analysis is accurate, well-supported, and ready to share or use for a decision. Use when reviewing methodology, calculations, comparisons, visuals, caveats, or conclusions.
Prepare KPI readouts, scorecards, WBR/MBR/QBR updates, and executive summaries from quantitative business or product metrics; use when the task is to report status, compare against targets, explain validated drivers, and state operating implications.
Analyze product or business data to support a decision or recommendation. Use when a decision depends on metric-backed evidence, such as choosing a direction, prioritizing an opportunity, evaluating a change, segmenting users, sizing tradeoffs, or deciding what to do next.
Design, build, revise, or QA quantitative charts and figures. Use when an analytical answer needs visual judgment, whether for an inline answer, report, dashboard, notebook, or artifact.
Audit or critique a product flow, journey, workflow, funnel, onboarding path, checkout path, settings path, screen, or multi-step product experience by capturing screenshots first, then reporting UX, design, and accessibility findings inline from that evidence. Use Figma only when the user explicitly asks for a board. Use when the user asks to audit, review, critique, inspect, assess, analyze, evaluate, or give feedback on a product experience.
Internal prototype QA helper. Use only after a Product Design prototype, URL-to-code build, or image-to-code build has a source visual target and a rendered implementation to compare before handoff. Do not use for broad UX critique, design critique, product audits, or flow reviews; route those user-facing requests to audit.
Mandatory design-brief gate for clarifying the product and outcome. Use before ideation, image-to-code builds, redesigns, or product UI work to clarify missing product information and play back the brief before proceeding.
Use when Product Design is explicitly invoked, or when the user's main goal is to explore a design, research UX, audit or critique a flow, faithfully clone a visual source, check a built design, or share a prototype. Do not use Product Design for ordinary implementation unless the user explicitly asks for it.
Generate image-based alternatives, remixes, or new design directions from a Product Design brief. Use when the user asks for design variants, visual exploration, remixes, or image-generated approaches from provided context.
Implement a selected image, screenshot, mockup, or Image Gen reference as a faithful, responsive frontend.
Create, edit, or validate reproducible SQL or Python notebooks. Use for notebooks, SQL/Python scratchpads, reproducible exploration, audit trails, or runnable companions where the analysis should be reviewable or rerunnable.
Run fast, source-grounded UX research on the highest-signal problems users are experiencing with a user-specified digital product. Use when the user asks to research user pain, UX friction, onboarding issues, docs/help problems, developer experience friction, support pain, product workflow issues, or current user complaints for a named product.
Share a runnable prototype using the user's preferred deployment tool.
Load or manage Product Design's saved user context. Use when the user asks to set up Product Design, get started, onboard, save product or design sources, see what Product Design remembers, update saved context, or remember Product Design preferences. Examples include product URLs, Figma files, screenshots, reference images, codebase paths, Storybook, tokens, design systems, brand assets, and general product/design notes.
Use when the user wants to know what changed with one account, monitor an owner portfolio or watchlist, or rank accounts needing attention from recent evidence. Produce an evidence-backed account brief or bounded watchlist summary with recommended actions.
Use when a small amount of missing context would materially improve the answer and tappable options plus a free-text answer can gather it efficiently.
Use only when a focused Sales workflow has selected a present and connected Apollo connector, or the user explicitly asks for Apollo prospecting, enrichment, Company Details, records, sequences, or outbound planning. Apply Apollo v2-specific behavior only after verifying app version 2.0.0 or later.
Review commercial proposals and build customer-led business cases, ROI or value models, pricing or investment rationales, executive summaries, and customer-ready value stories tied to a customer, workflow, initiative, or decision.
Use when the user wants a competitor or vendor comparison, market-landscape analysis, battlecard, objection package, positioning brief, or account-specific competitive view. Produce an evidence-backed comparison, guidance, and brief, using supplied materials, connected research, and public evidence when appropriate.
Use for data-first company, contact, lead, and prospect discovery or enrichment, including firmographic or technographic completion, named-company profiling, entity resolution, ICP matching, prospect-list building, segmentation, trigger or sales-signal analysis, market and territory scans, and enrichment-backed comparisons. Exclude meeting preparation, outreach-led company research, and prioritization of an already established account book.
Use when the user wants verbatim customer or prospect quotes, voice-of-customer evidence, theme validation, objection evidence, product-friction examples, or support for a product area, use case, or sales narrative. Retrieve from transcripts, call notes, supplied transcript-like recordings or exports, and other grounded call material using explicit speaker-confidence and provenance rules.
Use when a seller needs to find the best internal experts, owners, approvers, documents, channels, source-of-truth materials, or escalation routes for a customer question, product topic, competitive objection, implementation issue, account task, or other sales-support need. Route questions such as 'who knows about this?', 'what should I read?', 'where is the source of truth?', and 'which internal channel or document should I use?' here.
Use after a completed customer, prospect, partner, or important internal sales call, discovery session, demo, or conversation, including when the user supplies or uploads a transcript, notes, recording summary, or other grounded evidence. Advise on, assess, qualify, recap, and produce seller-ready follow-up actions and drafts.
Use only when a focused Sales workflow has selected a connected HubSpot CRM, or the user explicitly asks for HubSpot guidance, reads, drafts, notes, or reviewed record changes. Do not use when another CRM is authoritative.
Use this Sales index first for explicit Sales mentions and clear seller workflows: prospecting, lead qualification, account research, monitoring or prioritization, meeting prep, call follow-up, outreach research, deal strategy, pipeline or forecast review, CRM-backed context or data enrichment, internal source finding or sales support, customer quotes or evidence, business cases, competitive briefs, rep coaching, sales company research, and company or contact enrichment. For implicit use require clear seller, prospect, account, opportunity, pipeline, forecast, CRM, or customer-facing sales intent.
Use when the user wants evidence-backed coaching for one rep's calls, especially by comparing the rep with peer examples to identify repeatable best practices, specific upgrade moments, and practical next-call language.
Use when the user wants strategy for one active deal, renewal, negotiation, buying process, or initial sales motion for an offer or product. Build a grounded deal map or practical sales plan with objections, sequencing, posture, and prioritized next actions. Use prepare-for-meeting instead for multi-account same-day customer call queues.
Claude Skills — questions
Answers built from the skills we actually parsed.