Full index · agent skills
All agent skills — page 91 of 112
Every one of the 11,101 agent skills in the directory, ranked by adoption. Jump to a tool, or browse the agent skills hub.
- Remove Demo LimitsStrip Open Harness hosted-demo trial caps and deployment gates from your fork so self-hosted sessions and deletes work like a normal deployment.24 installs
- Safe Sql ExecutionImplement Supabase Studio–style SQL execution where only user-authored or hardcoded SQL runs, blocking attacker-influenced fragments from URLs or request bodies.24 installs
- Search Consumption CliLocate Microsoft Fabric lakehouse, warehouse, or notebook items by name when you do not know which workspace they live in, then pass IDs into downstream Fabric REST calls.24 installs
- Security Reviewer FormatsStandardize vulnerability-scan progress and final report blocks when orchestrating agent skill security assessments in audit, diff, or plan mode.24 installs
- Synapse MigrationPlan Azure Synapse Spark pool to Microsoft Fabric capacity SKU sizing when migrating analytics workloads to a Fabric workspace.24 installs
- Winui Code ReviewReview WinUI 3 / XAML apps against Microsoft performance, security, accessibility, globalization, and code-quality rules before merge or release.24 installs
- Winui Dev WorkflowFollow a structured WinUI desktop app workflow when building or iterating Windows client UI with Microsoft win-dev-skills.24 installs
- Flux 2 KleinGenerate fast Flux 2 Klein images on RunComfy with model-specific prompting, step-count, and 9B vs 4B routing instead of generic text-to-image guesses.21 installs
- Marketplace ScaffoldScaffold a Next.js Sitecore marketplace app with the right client-only or Auth0 full-stack template and registry packages.21 installs
- Cavekit Design SystemDefine and maintain DESIGN.md as the single visual contract so AI agents apply colors, typography, spacing, and components consistently across a Cavekit or agent-driven UI build.20 installs
- Cavekit MethodologyApply Cavekit specification-driven development—Hunt lifecycle and kits-first routing—so agents never skip from raw requirements straight to code.20 installs
- Cavekit RevisionStop one-off code hot-fixes from being erased on the next Cavekit iteration by tracing defects back to kits and prompts and fixing at the source.20 installs
- Cavekit Validation FirstDefine testable acceptance criteria and a six-gate validation pipeline so agent-generated work in Cavekit/SDD can be proven in CI instead of guessed.20 installs
- Marketplace Add ExtensionScaffold Sitecore Marketplace extension points—custom fields and dashboard widgets—with React/Next.js boilerplate wired to marketplace client hooks.20 installs
- Marketplace Build ComponentLook up and install Blok-themed shadcn UI blocks from the Sitecore marketplace into a React or Next.js app without guessing registry URLs.20 installs
- Marketplace Sdk ReferenceWire Sitecore Marketplace ClientSDK AI mutations such as brand review for copy and assets inside a marketplace-hosted app.20 installs
- Dynamic ResourcesCustomize pi-mono coding-agent interactive UI colors and message surfaces using the caveman-code dynamic-resources theme schema.19 installs
- Marketplace Add XmcImplement XM Cloud sites and related XMC ClientSDK queries in a Sitecore Marketplace React app with correct context IDs.19 installs
- Plugin CreatorScaffold a publishable Cave plugin with a valid `.cave-plugin/plugin.json` manifest and the standard commands, skills, agents, themes, and hooks folders before listing it on the marketplace.19 installs
- Agent Friendly ApisFix Next.js and REST API bugs that break agent callers—wrong status codes, broken filters, dynamic routes, and summary stats—using symptom-driven recipes.18 installs
- Better Sqlite3 RebuildRecover a local app that crashes on better-sqlite3 load after you change Node versions or pull fresh node_modules.18 installs
- Filesystem AgentsImplement filesystem-oriented agents with the AI SDK ToolLoopAgent pattern so the model can call tools in a loop until the task is done.18 installs
- Flux KontextEdit a single reference image with Flux 1 Kontext Pro on RunComfy using documented prompting patterns and the local RunComfy CLI.18 installs
- Gpt Image EditEdit product and marketing images with OpenAI GPT Image 2 on RunComfy while using bundled preservation, typography, and multi-reference prompting patterns.18 installs
- Design SystemKeep one DESIGN.md spec so every UI-building agent applies the same tokens, typography, and component rules instead of hardcoding visuals per file.17 installs
- Json Render React EmailTurn json-render JSON specs into HTML or plain-text transactional and marketing emails with @json-render/react-email and React Email components.17 installs
- Wan 2 7Generate Wan 2.7 text-to-video (lip-sync, multi-reference motion) on RunComfy via the local CLI with correct duration, resolution, and routing versus sibling models.17 installs
- Context ArchitectureStructure project context as a DAG with progressive disclosure so coding agents load only refs, kits, plans, and impl nodes relevant to each task.16 installs
- Convergence MonitoringKnow when to stop agent fix loops by watching shrinking diffs, pass rates, and ceiling signals instead of guessing iteration counts.16 installs
- Documentation InversionReplace rotting human wikis with CLAUDE.md, navigation skills, and plugins so agents load current, machine-readable docs from the repo.16 installs
- Driving Claude Code SessionsDrive long Claude Code sessions with checkpoints, handoffs, and context hygiene.16 installs
- Gpt Image 2Wire your agent workflow to OpenAI-style GPT image generation when you need illustrations, mocks, or marketing visuals from prompts.16 installs
- Happyhorse 1 0Generate 1080p text-to-video with synchronized audio through RunComfy CLI when HappyHorse 1.0 fits your prompt, duration, and aspect-ratio needs.16 installs
- Marketplace Add AiWire Sitecore Marketplace SDK AI modules into a React app so marketers can run brand-aligned text review inside the marketplace shell.16 installs
- Marketplace DeployDeploy a Next.js Sitecore marketplace app on Vercel with correct CSP and iframe embedding headers so Sitecore Cloud can embed the app without a blank frame.16 installs
- MethodologyRun the Cavekit Hunt lifecycle so specs and kits are defined before agents implement code, for greenfield builds or rewrites.16 installs
- Ui CraftImplement polished, accessible, performant UI—from components to landing pages—when you need implementation craft beyond a DESIGN.md spec.16 installs
- AiInstall this when you are building Python AI agents or chat apps and need the Vercel Python `ai` SDK patterns for models, streaming, tools, and structured output.15 installs
- Brownfield AdoptionLayer Cavekit kits onto a working legacy codebase using the documented brownfield process so agents gain traceability without a risky full rewrite.15 installs
- Cavekit WritingDefine implementation-agnostic Cavekit kits with testable acceptance criteria so Claude Code and similar agents can build or validate against durable WHAT-not-HOW requirements.15 installs
- Issue ManagementCreate and edit GitHub Issues 2.0 parent, sub-issue, blocked-by, and blocking relationships from the terminal using verified gh CLI flags.15 installs
- Peer ReviewRun a second model as a critical reviewer so your primary coding agent does not rubber-stamp its own output.15 installs
- RevisionInstall this when a Cavekit-driven build regressed or failed tests and you need to trace the defect to kits and prompts instead of patching code in isolation.15 installs
- Speculative PipelineOverlap multi-agent pipeline stages with staggered starts and convergence loops so long agent workflows finish in less wall-clock time.15 installs
- Impl TrackingMaintain a living implementation tracking document so each agent session knows what shipped, what is pending, and which approaches already failed.14 installs
- Prompt PipelineInstall this when you are designing Cavekit Hunt/SDD numbered prompt files that drive Spec, Plan, and Implement without bloating each prompt with content.14 installs
- Validation FirstDefine testable acceptance criteria and a six-gate pipeline so agent output can be automatically verified before you treat work as done.14 installs
- Zero NativeScaffold and ship a native desktop shell around a web UI using the zero-native Zig CLI, WebView or CEF, bridge commands, and packaging.13 installs
- Studio Mock Api TestsWrite Supabase Studio component tests that mock HTTP with MSW instead of vi.mocking data hooks.11 installs
- Peer Review LoopRun a Cavekit spec through a Ralph build loop where Codex adversarially reviews Claude’s implementation each iteration.10 installs
- Writing GuidelinesAudit docs and marketing prose against Vercel’s live Writing Guidelines with file:line findings.9 installs
- DeepsecRun deeper security checks in the dev3000 workflow when hardening a Vercel-oriented dev agent stack before ship.8 installs
- Pre ReleaseValidate Changesets in the Sandcastle repo are user-facing, grouped, and ready before you cut a release.8 installs
- Remotion Best PracticesEmbed Three.js / React Three Fiber 3D scenes in Remotion compositions without flicker during video render.8 installs
- What To TestDecide what to test in the ai-hero CLI and how to structure tests without chasing coverage for its own sake.8 installs
- Add Function ExamplesAdd or extend `examples/ai-functions` entries in the Vercel AI SDK repo to exercise new features against live provider APIs.7 installs
- AiWire up Python agents with the Vercel `ai` SDK—models, streaming events, tools, MCP, and structured outputs—for backend features in a solo project.6 installs
- Web Design GuidelinesAudit UI source files against Vercel’s live Web Interface Guidelines and get terse file:line findings before you ship or iterate on accessibility and UX.6 installs
- BackpropTurn test failures and bug reports into spec updates—§B backprop rows and §V invariants—so the same bug class cannot recur under spec-driven development.5 installs
- BuildRun a single-thread plan-and-execute loop against SPEC.md tasks, updating status cells and invoking backprop when verification fails.5 installs
- Caveman InternalShrink machine-to-machine agent handoffs and Cavekit artifacts so the next agent in the loop reads fewer tokens without touching user-facing replies.5 installs
- CheckRun a read-only SPEC.md versus codebase drift audit and get severity-grouped violations before you rewrite specs or schedule build work.5 installs
- Complexity DetectionScore a task on five axes so Cavekit can right-size budgets, model choice, and review depth before sketch, map, or make.5 installs
- Major Version ModeSwitch your coding agent into maintainer mode for the next Vercel AI SDK major release when you are intentionally shipping breaking API changes.5 installs
- Remotion Best PracticesInstall domain rules for Remotion so your agent scaffolds React video projects correctly and animates frames without CSS tricks that fail at render time.5 installs
- SpecMaintain a single repo-root SPEC.md (goal, constraints, invariants, tasks, bugs) before and during agent-driven builds.5 installs
- Write A PrdTurn a fuzzy feature idea into a scoped PRD filed as a GitHub issue before your agent starts coding.5 installs
- Automate Zero NativeInspect and drive a running zero-native WebView shell—list windows, snapshot state, reload, and capture screenshots—without manual clicking.4 installs
- Autonomous LoopUnderstand Cavekit's stop-hook autonomous loop, budgets, and completion sentinel before running /ck:make or debugging sessions that will not stop.4 installs
- Bailian Model RecommendPick the right Alibaba Cloud Bailian model for your AI feature before you wire up API calls or prototypes.4 installs
- Capability DiscoveryScan the real environment for CLIs, MCP servers, and Claude plugins before Cavekit kits assume tools that are not installed.4 installs
- Graphify IntegrationGive Cavekit architect, researcher, reviewer, and task-builder agents symbol-level dependency queries instead of blind grep when refactoring or assessing blast radius.4 installs
- Happyhorse Prompt StudioTurn a vague “I want a short video” idea into a HappyHorse 1.0 prompt with scene, motion, audio, and reference-image syntax.4 installs
- Karpathy GuardrailsLoad four behavioral rules so Cavekit agents think before coding, avoid scope creep, and make surgical, goal-driven edits on every task.4 installs
- Tw BlocksCopy proven TW-Blocks spacing patterns when you build dashboards, sidebars, forms, and data-heavy SaaS screens without guessing padding and grid rhythm.4 installs
- Vercel React Best PracticesGive your agent a prioritized React and Next.js performance rulebook while you generate, refactor, or review UI code.4 installs
- Agentmemory AgentsRun agentmemory connect for your host (Claude Code, Cursor, Codex, Cline, Copilot CLI, etc.) so MCP points at persistent memory without hand-editing JSON.3 installs
- Agentmemory ArchitectureUnderstand how agentmemory stores, retrieves, and serves agent observations so you can extend or debug your local memory server.3 installs
- Agentmemory ConfigLook up all 34 AGENTMEMORY_* environment variables and where agentmemory reads config from ~/.agentmemory/.env.3 installs
- Agentmemory HooksUnderstand and debug how the agentmemory Claude Code plugin auto-captures session memory via lifecycle hooks without manual saves.3 installs
- Agentmemory Mcp ToolsLook up the 53 agentmemory MCP tools, core subset, and parameters when wiring persistent agent memory in Claude Code or Cursor.3 installs
- Agentmemory Rest ApiLook up every AgentMemory REST path and auth pattern when wiring your agent stack to the local memory server on port 3111.3 installs
- DeeppapernoteTurn one paper (DOI, arXiv, PDF, or Zotero item) into a structured, evidence-based deep-reading Markdown note for your vault or workspace.3 installs
- Finops Duizhang ExpertApply FinOps-style cloud billing reconciliation (对账) workflows when monthly spend must match invoices and allocation tags.3 installs
- Implicit CadGenerate browser-rendered implicit CAD models as `.implicit.js` files using GLSL signed-distance primitives and snapshot exports for quick physical or visual prototypes.3 installs
- Llm WikiLet your coding agent query pages, files, and the wikilink graph in the locally running LLM Wiki desktop app over its HTTP API.3 installs
- Styled JsxScope React and Next.js component styles with styled-jsx without global CSS leaks or class-name collisions.3 installs
- Vercel Composition PatternsRefactor or generate React UI using Vercel-style composition—compound components, lifted providers, and explicit variants instead of boolean prop sprawl.3 installs
- Write Agentmemory SkillAuthor or refactor agentmemory plugin skills so they follow the house SKILL.md tier, triggers, and cross-link rules before contributors merge.3 installs
- Agent TeamsAssign conjure agent-team members to implementer, researcher, tester, reviewer, or architect roles with a clear capability matrix before parallel work starts.2 installs
- Brainstorming Ideas Into DesignsClaude Code agent workflow helper from OBRA clank repository.2 installs
- Code Review ReceptionClaude Code agent workflow helper from OBRA clank repository.2 installs
- Commenting IntentClaude Code agent workflow helper from OBRA clank repository.2 installs
- Condition Based WaitingClaude Code agent workflow helper from OBRA clank repository.2 installs
- CoolifyAutomate infrastructure deployments and manage applications, databases, and services through the Coolify REST API.2 installs
- Creating SkillsClaude Code agent workflow helper from OBRA clank repository.2 installs
- Defense In Depth ValidationClaude Code agent workflow helper from OBRA clank repository.2 installs
- Designing Before CodingClaude Code agent workflow helper from OBRA clank repository.2 installs
- Dispatching Parallel AgentsClaude Code agent workflow helper from OBRA clank repository.2 installs
- Domain Focused NamingClaude Code agent workflow helper from OBRA clank repository.2 installs