amazon-push-score quantifies amazon listing performance via a 4-dimensional weighted model (ctr, cvr, 7-day growth, account health) to classify traffic tier and generate optimization priorities. calculation is deterministic but assumes metric accuracy from user input.

kids-points-v2 is a feishu-native family points ledger that parses natural language chinese via llm, persists to sqlite, and exposes cli commands for scripting. v2 upgrades v1 with concurrent-safe storage, group chat support, and deduplication.

linkfox-tsearch-search performs web searches with built-in content extraction, returning unstructured text from top results. covers news, product research, reddit discussions, trending topics, and real-time data lookup with keyword-based querying and inline summarization.

ai-ges-huawei-cloud-ges-graph provides cypher and gql query execution against huawei cloud graph engine service, covering node/edge operations, schema management, and import/export via python or node.js scripts with aksk or token authentication.

vault-sync-engine addresses obsidian multi-device sync via git, icloud, dropbox, or syncthing. handles conflict detection and three-way merge with selective sync rules for .obsidian/ configs and health monitoring.

aggregates security advisories from multiple sources and ranks alerts by asset relevance and exploitability, reducing noise from hundreds of daily advisories to actionable findings. includes offline caching, incremental deduplication, and tiered notification strategy.

musescore-mcp bridges Claude to MuseScore's web interface via a browser extension, enabling score search, metadata retrieval, and PDF conversion without direct API access. read-only, authentication routed through your signed-in session.

hsk-learning tracks mastery of 2,211 chinese characters via spaced repetition, parses quiz logs to update proficiency, and generates adaptive quizzes filtered by hsk level and review status.

oo-dynatrace wraps the dynatrace connector via oo CLI for entity and problem queries. reads only (get, list operations) with server-side credential injection. no write or destructive actions documented.

agentic-video-editor automates mp4 editing via natural language prompts, handling upload, ai-driven cuts and transitions, and cloud rendering. users describe desired output and receive polished 1080p exports without manual intervention.

toingg-skill bundles campaign creation, optional daily analytics scheduling, and excel-to-whatsapp contact broadcast via five python helpers sharing a bearer token. covers setup, three main workflows, and file mapping but lacks explicit failure modes.

dlazy-recraft-v4 wraps the dLazy hosted API for 1MP image generation via CLI. authentication is device-code flow or manual key entry. supports aspect ratio control and async task polling. returns PNG URLs hosted on dLazy storage.

video-editor-clips automates ai-driven video trimming, transitions, and beat-sync rendering via cloud gpu. accepts mp4/mov/avi/webm up to 500mb and returns edited exports in 30-90 seconds without local software installation.

make-textbook orchestrates web research into structured graduate-level textbooks as formatted PDFs, with defined passes for outline approval, chapter research, content drafting, QA, and generation. the workflow is explicit but execution details remain vague.

capacity-planning produces a comprehensive capacity document with baseline metrics, growth projections, resource requirements, scaling strategy, cost forecasts, and infrastructure roadmap. scope is broad but execution relies heavily on user-supplied data; lacks concrete vendor-specific trigger phrases.

financial-statement-explainer translates P&L, balance sheet, and cash flow statements into plain english for non-accountants. it walks through structure section-by-section, identifies key ratios, and surfaces the business narrative without inventing missing data.

release-guard performs local, non-destructive pre-publish checks on openclaw skill folders, flagging file counts, secrets, unrelated workspace artifacts, and risky shell patterns without executing code or contacting external services.

hcc-meta-ai provides a six-element prompt framework for ai-assisted meta-analysis of hepatocellular carcinoma treatments, spanning literature screening through outcome extraction. validates >95% accuracy on dual independent tasks across three language models.

linux-security-scanner audits ssh config, open ports, firewall rules, failed logins, sudoers permissions, world-writable files, and suid binaries via a single portable shell script with modular check options and colorized reporting.

free-image-maker generates promotional images and videos from text prompts or uploaded files via a remote gpu pipeline. users describe what they want and receive rendered output within 20-90 seconds. requires nemo token authentication and credit balance.

agent-fuel manages autonomous wallet balances with moonpay top-ups and x402 payment handling. enforces daily spend limits and alerts humans on threshold breaches.

agent-pipeline orchestrates code task workflows by sequencing specialized sub-agents for research, implementation, review, security audit, and testing, with board logging and planner integration.

ai-video-editor-completely-free wraps a nemo video cloud api for trimming, transition insertion, and export to mp4/mov/avi/webm. requires nemo token auth, session creation, and multipart upload. processes clips in 1-2 minutes at 1080p max.

dbskiter-db-scheduler covers database backups, cron-based task scheduling, and dag workflow orchestration with log inspection. supports full/incremental backups, task enable/disable, and daemon process management.

jd-link-converter transforms jingdong mobile share links and mobile product urls into clean pc-endpoint links. handles short link resolution via http redirect tracing, mobile-to-pc url normalization, and batch conversion with fallback regex extraction.

abm-outbound chains linkedin scraping, apollo enrichment, skip trace address lookup, and multi-channel delivery (email, linkedin, handwritten mail) into a coordinated outreach pipeline. solid procedural walkthrough but lacks failure handling and specific trigger patterns.

md-converter transforms markdown files to docx, pdf, and html with chinese font support. includes setup instructions, format-specific command routing, and a feature compatibility matrix.

caption-burner accepts video files and natural language instructions to bake hardcoded captions into mp4 exports via remote gpu processing, targeting creators who need platform-agnostic subtitled video without manual timeline work.

mindmap-creation structures knowledge hierarchically using mece principles and branching rules. produces markdown mind maps from topics without commentary, suitable for study notes and project planning.

oo-anchor-browser wraps anchor browser operations through the oomol cli, supporting billing queries, project metadata retrieval, and session startup. credentials are server-injected; no raw token handling required.

windows-control provides low-level mouse, keyboard, and window automation primitives for windows applications. it lacks a coherent intent/decision model and treats skill invocation as direct script calls rather than declarative actions.

video-editing-ai-by-prompt accepts raw video uploads and natural language prompts to generate edited 1080p mp4 files via a cloud gpu pipeline. users describe edits (cut pauses, add music, overlay text) and receive finished video within 30-90 seconds without local software.

civitai-api-art wraps the civitai javascript sdk to generate images from prompts with configurable models, samplers, lora networks, and generation parameters via command-line options.

stockstar-esg-skill queries listed company esg ratings from three rating agencies (miotech, chindices, syntaogf) covering a-shares and hong kong stocks. supports single stock lookup, detailed scoring, multi-stock comparison, and rating lists via python cli wrapper.

desktop-notify triggers system notifications and sounds across windows, macos, and linux when long-running tasks complete. supports custom titles and message text via script parameters or global setup mode that auto-enables per-conversation reminders.

unifi-cli queries a local unifi gateway to answer questions about network hardware, connected devices, access points, and gateway status via the unifi cli tool.

agenthire enables agents to search and hire specialized ai agents on base sepolia testnet with automatic escrow-based payment. agents can delegate tasks like token swaps, research, and translations to external providers and verify results on-chain.

skill-gws wraps the official google workspace cli, exposing drive, gmail, calendar, sheets, docs, chat and admin apis. includes mcp server mode for ai agents and dynamic api coverage via google discovery service.

feishu-voice-send converts text to native feishu voice messages via minimax or edge tts, with automatic fallback on quota exhaustion. receives and transcribes voice via whisper, handling multi-language input.

video-open-ai automates video editing via cloud GPU rendering. users upload clips and describe edits in natural language, receiving 1080p mp4 output within 1-2 minutes without manual timeline work.

autoskill analyzes problem statements and contexts to score available skills for relevance, recommending the most applicable ones in priority order with mandatory user confirmation before any execution.

p-image-edit performs reference-driven image composition and inpainting via async http polling, with optional upscaling and chaining to video pipelines. covers 1-5 image inputs, aspect ratio control, and turbo mode toggling.

skill-install-bridge generates standardized install commands, readme snippets, and social copy for published clawhub skills, with output validation and overwrite safeguards.

conversation-video renders multi-speaker transcripts into animated videos with synced audio and text overlays, supporting both fast ffmpeg rendering and feature-rich remotion-based workflows.

database-migration-risk-review classifies schema changes and production hazards (locks, compatibility, backfills, rollback), proposes expand-contract sequencing for risky operations, and defines verification checkpoints. works standalone with provided code.

justoneapi-bilibili-get-video-caption wraps the bilibili caption endpoint for transcript extraction and multilingual analysis, requiring aid, bvid, and cid parameters passed as query strings through justoneapi's v2 api.

openclaw-create-feishu-agent-skills automates single-pass agent creation and config wiring for feishu via routing mode selection, cli invocation, and json mutation with checkpoint validation.