ai-rag-pipeline sketches retrieval augmented generation via belt CLI but omits essential implementation detail, decision logic, and failure handling, leaving adopters to reverse-engineer the remaining 80 percent.

dbs-knowledge is a folder-to-knowledge-base conversion tool that scans local file structures, generates navigation files (SOURCE_OF_TRUTH.md), handles file ingestion, retrieval and health checks. it avoids external databases and preserves original files unchanged.

multi-search-engine aggregates queries across 17 search providers (baidu, google, duckduckgo, etc) with regional filtering for cn and global results. configuration-driven url templating.

game-ui-design provides design philosophy and principles for shipping game interfaces across platforms, but lacks actionable procedures, specific trigger scenarios, and concrete decision workflows.

file-organizer identifies trigger scenarios for folder cleanup but lacks actionable procedure, decision logic, and failure handling. primary value is use-case enumeration rather than implementation guidance.

landing-page-design documents ai image generation for hero sections via belt cli, but lacks conversion mechanics, layout guidance, cta strategy, and mobile considerations promised in the description.

ai-video-generation wraps inference.sh cli to invoke 40+ video models including veo and seedance, but lacks procedural depth, error handling guidance, and clear intent boundaries.

allium is a domain-level behaviour specification language positioned between feature descriptions and implementation. it generates integration tests and forces ambiguities into the open before coding begins.

axiom-sre establishes sre principles and cognitive discipline for incident response, but lacks concrete procedures, trigger phrases, and actionable workflows for agents to execute.

tiktok-video-downloader is a browser extension for saving tiktok videos as mp4 files. lacks actionable procedural detail, edge case handling, and concrete trigger guidance for ai agent deployment.

code-refactoring-context-restore addresses semantic memory rehydration for multi-agent ai workflows, offering vector search and context ranking strategies. lacks concrete refactoring procedures and actionable recovery steps.

pulumi-arm-to-pulumi covers arm template conversion and optional resource import workflows. emphasizes manual translation, zero-diff validation on import, and formal reporting but lacks concrete examples and specific trigger guidance.

wterm provides DOM-based terminal emulation with zig/wasm parsing and react bindings, but documentation is incomplete - no usage procedures, error handling, or deployment specifics.

media-use orchestrates asset resolution across bgm, sfx, image, icon, voice, and avatar types through heygen cli binding. scope unclear, procedure incomplete, edge cases absent.

react-native-best-practices is a routing document listing eight sub-skills covering animations, gestures, svg, on-device ai, rich text, multithreading, audio, and jsi. it contains no actionable guidance, procedures, or decision logic.

story-long-scan provides ranking analysis across chinese web fiction platforms (qidian, fanqie, jjwxc, qimao) with platform-specific scripts and data extraction workflows, but the actual analytical procedures and decision logic remain underspecified.

aicoin-account provides exchange account read-only operations including balance, positions, and order history queries, but the documented skill definition is incomplete and lacks actionable procedure detail.

error-diagnostics-error-trace provides high-level guidance on error tracking and observability concepts but lacks concrete procedural steps, specific tool integration paths, or actionable error-handling workflows.

after-effects covers disney's 12 animation principles with partial implementation guidance, mixing code snippets and manual techniques without complete workflow coverage or decision logic.

personal-finance-coach declares expertise in tax optimization and investment theory but provides no actionable procedures, decision logic, or concrete guardrails. the skill.md is a header with dependencies and no operational content.

landing-page-design attempts to cover conversion optimization through ai image generation via a belt cli wrapper, but lacks coherent guidance on layout principles, cta mechanics, or design decision-making.

clinical-reports outlines regulatory principles for healthcare documentation across case reports, diagnostics, and trials, but lacks actionable procedures, specific triggers, and failure mode handling required for consistent skill execution.

circle-downloader is a browser extension that extracts videos from circle.so communities and embedded platforms (loom, vimeo, youtube, wistia, tella) and saves them as mp4 files locally. includes 3 free downloads.

expo-dom bridges web libraries to native by rendering React web code in webviews on ios/android while preserving web semantics. enables incremental migration of existing web components without rewriting.

toolkit for roblox murder mystery 2 players to track inventory, analyze gameplay statistics, and generate performance reports through command-line and python api interfaces.

mm2-roblox-analytics-toolkit claims to provide murder mystery 2 game analytics and inventory tracking, but the skill.md cuts off mid-setup without defining intents, inputs, outputs, or actual workflow beyond installation commands.

cut-the-curve catalogs seven motion-transition techniques (zoom-through, inverse zoom, curve cuts, waterfall, rack-focus, entry, nudge) with axes and use cases, but omits procedural steps, inputs, and decision logic.

update-docs orchestrates multi-agent documentation workflows for local code changes, covering markdown files, readmes, and api specs, but lacks detailed procedural definition.

notion-cli wraps the notion api as a single go binary with json and table output modes. installation instructions provided, but skill lacks operational guidance.

tavily-crawl extracts multi-page website content via the tavily cli, with local markdown output. incomplete documentation and vague procedures limit practical applicability.

wistia-video-downloader is a browser extension that extracts and saves wistia-hosted videos as mp4 files from embedded players and business pages, automating media source detection for offline access.

ljg-card casts content into png visuals across seven mold types (long reading, infograph, multi-card, visual notes, manga, whiteboard, large text). outputs only, no procedural flow defined.

extract-errors scans react source for error messages and assigns unique codes, detecting unassigned errors via a yarn command. synchronization validation is minimal.

gemini-api lists enterprise google ai capabilities (text, multimodal, function calling, embeddings, batch prediction) but lacks actionable procedures, failure modes, or working examples for agent platform integration.

onboard-context-matic provides a partial guided walkthrough for mcp server onboarding. the document establishes conduct rules but leaves procedural phases incomplete and fails to specify clear outcomes or failure modes.

solve-challenge claims to triage ctf challenges and route to specialized skills, but provides only environment setup instructions without actual triage logic, routing mechanism, or decision framework.

audit is a diagnostic framework for measuring code quality across five dimensions (accessibility, performance, theming, responsive design, anti-patterns). it documents issues without fixing them, requires upstream context gathering, and produces severity-tagged findings with remediation guidance.

multi-repo-swarm describes cross-repository orchestration via gh cli and a fictional ruv-swarm npm package, with bash examples for dependency updates and pr creation across repos. concrete procedures exist but rely on undocumented tooling.

sub-account wraps binance's master-account api endpoints for managing virtual sub-accounts, transfers, futures positions, and managed portfolios. endpoint reference provided but no procedural guidance for common workflows.

self-improvement prescribes a markdown-based logging system for agents to capture errors, learnings, and feature requests in project .learnings/ directories. promotes entries to shared memory files when broadly applicable.

landing-page-design covers ai-generated visual assets for landing pages via inference.sh cli, but lacks substantive guidance on conversion optimization, layout principles, or cta strategy despite claiming it in the description.

agent-performance-optimizer claims capability in bottleneck identification and resource analysis, but lacks concrete procedures, decision logic, input/output contracts, or failure recovery patterns.

langsmith-evaluator exposes environment variable setup and a single CLI command for listing evaluators, but omits intent, procedure steps, decision logic, and output contracts entirely.

documents assistant-ui tool mechanics including request-response and bidirectional streaming patterns, tool result caching, and reference links to implementation details, but lacks concrete trigger scenarios or actionable implementation guidance.

design-is orchestrates dieter rams audits by delegating evidence collection to subagents, scoring ten principles, and handing off verdicts to /make-plan. comprehensive framework but abstract and untested against real designs.

vitest is a vite-powered test framework with jest compatibility, but this skill document reads as a feature list rather than an actionable procedure for agents to write or run tests.

agent-coordination provides command syntax and agent type taxonomy for spawning multi-agent systems, but lacks executable procedures, failure modes, and concrete coordination logic.