searxng-search-skill wraps a python library for privacy-focused metasearch against local or remote searxng instances. includes retry logic, timeout handling, category and time filtering, and engine selection. mostly code samples without clear intent mapping or decision procedures.

claw-planops parses project notifications into structured tasks, generates micro-task plans and calendar exports (ics format), checks file-based progress evidence, and recommends dynamic rescheduling. targets chinese-language competition and delivery workflows.

hd-generator-no converts video clips to 1080p mp4 files via a cloud gpu pipeline. users upload files, describe desired output, and receive processed video without watermarks. requires nemo token authentication and processes most formats up to 500mb.

nexustrader bridges to a local mcp server for querying balances, positions, and tickers across binance, bybit, okx, bitget, hyperliquid. order placement requires explicit user confirmation.

poyo-claude-opus-4-7 wraps the claude opus 4.7 messages api endpoint on poyo.ai, covering payload construction, streaming setup, tool definitions, structured output, and prompt caching for server-side integrations.

lianjia covers second-hand home transactions, rental agreements, and tax estimation for beike ecosystem users. provides listing literacy, workflow checkpoints, and common pitfalls but lacks procedural depth on actual agent interaction patterns.

vn-video-editor accepts raw smartphone clips, applies ai-driven edits via cloud gpu rendering, and exports 1080p mp4 within 1-2 minutes. user describes desired output in natural language without timeline interaction.

supervision-layer enforces non-bypassable tool call guardrails including timeouts, circuit breakers, audit logging, and crash loop detection across agent invocations.

settle-agent-invoice handles multi-step invoice settlement via cai transfer after user confirmation of payee and amount, including recipient resolution and transfer status proof.

propai-sync covers monorepo maintenance, hosted platform validation via smoke tests, railway deployment checks, and clawhub skill publishing with ordered endpoint verification and handoff documentation.

rapport-memories indexes agent memory files and session transcripts into a vector store, enabling semantic search over past conversations and decisions to maintain contextual awareness across multi-turn interactions.

cms-report-issue handles the complete lifecycle of skill problem feedback: submitting bug reports and improvement suggestions, listing issues with statistics, and marking them resolved or closed. supports stdin piping of error output.

three-line-dragon-stock identifies momentum stocks via moving average alignment (ma5 > ma10 > ma20), volume confirmation, and six entry signals. targets limit-ups and platform breakouts with position sizing and stop-loss rules tied to entry method.

poyo-seedream-4-5 wraps the PoYo Seedream 4.5 api for text-to-image generation and image editing at 2K/4K resolutions, supporting multi-reference workflows and configurable output counts up to 15 images.

travel-experience-peak-end-design applies peak-end rule psychology to itinerary planning, redirecting budget concentration toward one engineered emotional peak and a strong closing moment to drive referrals and repeat bookings.

ria-position-sizing-kelly applies fractional kelly framework to cap single-name and concentrated positions within ipa bounds, treating ruin risk as the binding constraint for household capital allocation.

video-clip-maker-online-free enables remote gpu-backed video editing for trimming, transitions, and export to mp4. supports mp4, mov, avi, webm files up to 500mb with 30-90 second render times. routes user intent through keyword classification to appropriate api endpoints.

iflytek-speed-transcription converts audio files (wav, pcm, mp3) to text via the iflytek api with support for 202+ chinese dialects, automatic language detection, and optional speaker separation. transcription speed reaches approximately 20 seconds per hour of audio.

linkedin-person-search queries upkuajing's linkedin dataset by name, company, position, and industry. covers talent acquisition workflows with api key management, fee notification, and task continuation for large result sets.

talking-head-recut layers timed graphic cards (titles, lower-thirds, data callouts, quotes) onto existing video, synced to transcript. the agent designs each card as html, assembles them into a composition, and renders to mp4 via hyperframes.

veo-3-1 wraps poyo ai's frame-conditioned video generation api, supporting fast and quality modes with start/end frame guidance, up to 4k resolution, and reference image inputs via task submission and polling.

seedance-2 wraps poyo ai's image-to-video api for submitting jobs at 4s, 8s, or 12s durations with optional audio and fixed-lens control. covers model selection, request formatting, and task id retrieval.

meme-signal aggregates four free dex and trending data sources to score meme tokens 0-100 and flag risk tiers. workflow runs every 5 minutes, pulling from dexscreener, geckoterminal, pump.fun, and coingecko with narrative tagging.

auto-cleaning-disk removes temp files, browser cache, trash, and logs across windows, linux, and macos. offers auto or confirm mode. delegates os-specific logic to external reference files without detailing actual implementation.

browser-auto-plus provides multi-browser automation with retry logic, screenshot verification, and web scraping. lacks explicit intent definitions and concrete decision trees for when to failover or escalate.

cms-meeting-materials fetches raw meeting transcripts from cms ai huiji via pull apis, supporting list queries, full retrieval of completed meetings, and incremental polling of in-progress sessions. analysis and presentation remain the responsibility of consuming applications.

queries imdb title redux overview data via justoneapi, accepting an imdb id and optional language-country locale to retrieve concise summary and key attributes for movies or tv shows.

rss-content-flow monitors multiple rss feeds, extracts recent articles via structured parsing, and filters by age and content quality. outputs json or formatted lists for downstream processing by content generation skills.

paper-recommendation automates arxiv discovery, parallel sub-agent review, and structured briefing generation with daily cron scheduling. covers fetch, pdf extraction, and markdown output but lacks error handling and assumes human decision-making at critical junctures.

podcast-to-substack coordinates rss and notion ingestion to publish podcast episodes on substack with apple podcasts embeds, then generates linkedin companion posts from unified content.

youtube-creator automates video editing workflows by accepting raw footage and processing instructions via cloud GPU rendering, returning polished mp4 exports in 1-2 minutes without local tooling.

volcengine digital human video generator automates avatar creation from photos and scripts, handling gender detection, tts voice matching, and video synthesis via volcengine apis.

client-onboarding-automator sequences intake, proposal, contract, and payment collection into a six-step workflow with templated emails and stripe integration. covers project folder creation and timed welcome messages but lacks error handling.

mc-db-discovery provides a linear troubleshooting path for locating mission control's sqlite database and resolving schema mismatches between frontend interfaces and backend storage when standard paths fail.

chis-chisf provides lightweight aliases for openclaw skill installation with force-install variants and workspace-aware path handling, targeting repetitive install operations in a specific user environment.

image-analyzer-advisor structures visual analysis into five categories (elements, composition, quality, insights, recommendations) but lacks procedural rigor around tool invocation, error handling, and output validation.

clawdvine-skill-2 generates short-form videos for agents using various ai models with usdc payment on base via x402 protocol. no signup required, but agents can join the network for credits and onchain identity via erc8004.

task-management provides a framework for decomposing complex work into subtasks, tracking execution state through pending/in-progress/completed/deleted transitions, and managing dependencies. lacks concrete procedures for the stated 5-step flow and minimal failure-mode guidance.

apidot-gpt-5-5-api routes gpt 5.5 integration questions to apidot docs and model pages. documentation-only skill covering chat workflows, agent planning, request validation, and security practices for openai-compatible premium model access.

product-launch-video orchestrates a six-step workflow from brief through final render, routing product marketing requests into a time-coded shot sequence architecture. it pairs explicit procedural gates with references to external skill docs and preset systems, though execution relies heavily on undocumented helper scripts and assumes deep familiarity with the hyperframes ecosystem.

lessac-offline-voice-system recovers telegram voice messaging in openclaw by restoring faster-whisper transcription, configuring microsoft tts provider, and enabling the bundled speech plugin after upgrades or failed deployments.

google-my-business provides cli-driven integration for managing locations, questions, answers, and place action links via the membrane platform. authentication and credential refresh are handled automatically. discovery and execution of actions follows a search-then-run pattern.

fidel skill enables card transaction data access and financial workflow automation via membrane cli. covers authentication, connection setup, action discovery, and direct api proxying for fintech integrations.

toji-google-places wraps the google places api for location and business queries, returning ratings, hours, contact info and directions via command-line scripts. covers search, details, and nearby radius operations.

video-to-subtitle-generator automates subtitle creation and video rendering via remote gpu nodes. users upload files, describe desired output (language, format, burns), and receive processed 1080p video within 30-90 seconds. requires nemo token and api session setup.

per-agent-compression-universal auto-discovers openclaw agents and schedules weekly memory consolidation tasks that extract factual content from daily notes into domain-specific long-term files with deduplication and state tracking.

fathom connects to fathom ai for retrieving call recordings, transcripts, and summaries. supports filtering by date, speaker, and text search. requires api key setup and curl/jq.