blueai-nano-banana wraps gemini/nano-banana image generation via internal relay, outputting png locally. intended for rapid prototyping (moodboards, demo placeholders, social assets), not production design work.

cross-chain-arbitrage-cn automates discovery and execution of price discrepancies for tokens across 11 uniswap-supported chains, handling scanning, cost calculation, risk gating, and erc-7683 settlement. supports scan-only mode for research without capital deployment.

keplerjai-image-gen wraps three image generation models (gemini, minimax, seedream) via thinkzone ai. requires user model selection when unspecified, exposes cli scripts with model-specific parameter tables, but lacks explicit failure handling and recovery paths.

astock-daily-brief collects chinese stock market data via web search or optional api, assembles it into a structured daily report with indices, sector rankings, capital flows, and sentiment metrics. bilingual (english and simplified chinese) documentation.

aegean-cron-mastery covers scheduling reliable reminders and maintenance jobs in openclaw 2026.2.15+, distinguishing heartbeat drift from cron precision. includes payload patterns for push notifications vs silent logs and legacy migration guidance.

mu-pet is a macOS-only pixel art desktop overlay that roams the screen and responds to agent state changes via local HTTP API. the pet animates through nine states including idle, walking, working, thinking, and talking, with a right-click menu for manual control.

hyperliquid-claw provides command-line trading and portfolio monitoring for hyperliquid perpetual futures, covering balance checks, position tracking, order execution, and market analysis via coingecko data and momentum signals.

aura-alert-listener provides a lightweight polling loop with deduplication and conditional escalation for high-frequency alert checking. designed to replace frequent cron jobs and suppress noisy no-op cycles via state files and isolated sessions.

lovefromio-garmin-health-analysis exposes garmin connect metrics (sleep, body battery, hrv, heart rate, stress) via command-line scripts and generates interactive html dashboards. natural language queries map to data fetches and chart rendering.

kadence-theme-design-system provides WordPress templating via Kadence Theme with explicit Free vs Pro capability gating. covers design tokens, layout patterns, and template architecture with a pre-flight capability check to prevent invalid implementations.

clawme sends browser instructions via chrome extension, executing form fills, tweets, emails, clicks, and extracts with user confirmation. runs in authenticated user sessions unlike headless browsers.

video-narration adds ai voiceover to video clips (mp4, mov, avi, webm up to 500mb) via cloud gpu processing. renders 1080p output in 30-90 seconds. requires nemo token and session management for content creators and marketers.

fhe-service wraps a clinical scoring api that computes 165 rheumatology and general health metrics on fully homomorphic encrypted patient data. supports batch computation, per-score micropayments in usdc, and makes no plaintext logs of patient values.

qwen-code-skill wraps alibaba cloud qwen code cli for background task execution, code review, and automation. integrates with openclaw process management for session tracking and headless ci-cd workflows.

heyvideogen combines heygen hyperframes html animation with minimax hailuo ai video generation for vertical short-form content. targets users seeking simpler html-based workflows over remotion, with three content modes (educational, narrative-driven, drama) and full ffmpeg rendering pipeline.

architect-review evaluates system structure against openspec requirements, producing multi-dimension scorecards with actionable architectural recommendations. skips code-level detail in favor of module boundaries, data flow, and design trade-offs.

grit-the-power-of-passion-and-perseverance teaches angela duckworth's framework for sustained achievement through passion and perseverance. covers effort compounding, deliberate practice, passion cultivation, and hope as the engine of long-term goal pursuit.

converts user photos (portraits, pets, objects, logos) into four-action q-version gif sticker packs with chinese captions via static image generation, video animation, and format conversion with online preview and download capability.

amernet-ai-saas routes messages from multiple messaging channels (whatsapp, telegram, slack, discord, imessage) to an ai saas agent backend, maintaining per-user conversation context and supporting conversation reset and status checks.

managed-agent-growth-loop defines a repeatable operations framework for customer-resident agents, covering state handoff, proactive watchlists, team onboarding gaps, and outcome feedback. execution is strong but depends on manual discipline to maintain state and close feedback loops.

smyx-plant-species-recognition-analysis identifies plant species from images using deep learning and computer vision, outputting structured data including scientific name, family, growth habits, and care guidance. supports local files and urls with automatic history report retrieval from cloud api.

qq-mail-manager covers qq mailbox operations including retrieval, search, deletion, and sending. it enforces strict folder-plus-id localization, requires user confirmation before destructive actions, and uses single-page pagination with explicit next-page prompts.

typeform skill proxies the typeform api through maton's oauth gateway, covering form crud, response retrieval, and insights. requires managed connection setup and maton api key.

qdro-drafter converts divorce-decree retirement provisions into plan-typed draft court orders (qdro, coap, mpdo, rbco, governmental dro) with pre-approval routing to plan administrators and explicit survivor-benefit elections, flagging timing risks and pii boundaries.

etix-mcp wraps etix event search, venue lookup, and detail retrieval via mcp tools, routing through your signed-in browser tab to bypass datadome bot detection. no account required for public event discovery.

cruit-candidate guides users through setting up and maintaining a Cruit hiring profile for ai-native developers. includes safety rules, 7-step onboarding, profile drafting, and optional weekly refresh automation.

publish-to-wechat converts markdown to wechat-formatted rich text and publishes via wenyan cli, with automatic image upload to media library and yaml frontmatter-driven metadata. covers basic troubleshooting but lacks concrete decision logic for agents.

apidot-meshy-6-3d-api routes meshy 6 3d generation workflows through apidot, covering text-to-3d, image-to-3d, and multi-image modes with async task submission and webhook patterns. documentation-only skill with no executable code.

entry point skill for strategic game classification that diagnoses game type (sequential vs. simultaneous, zero-sum vs. non-zero-sum) and routes users to specialized game-theory skills. covers named game patterns including prisoners' dilemma, chicken, coordination, and assurance games.

bb-browser-claw wraps chromium automation via chrome devtools protocol, exposing 103 commands across 36 platforms through named adapters (twitter, weibo, bilibili, etc) and raw browser control. real sessions and cookies work without api keys.

converts still images into animated video clips via cloud gpu rendering. supports jpg, png, webp, bmp up to 200mb, returns 1080p mp4 within 30-90 seconds. requires nemo_token authentication and tracks session state across edit cycles.

image-to-online converts photo batches into 1080p shareable videos via cloud gpu rendering. accepts jpg/png/webp/heic up to 200mb, returns mp4 within 30-90 seconds. requires nemo_token authentication and session management.

search douyin users by keyword through justoneapi with optional pagination and user type filtering. returns profile metadata and follower signals for creator discovery.

intellectia-stock-forecast-2 wraps skillboss api hub to deliver yearly stock price predictions (2026-2035) and buy/sell/hold recommendations via web search and llm chat endpoints. supports single-symbol queries with structured output.

video-editor-inshot wraps a cloud gpu rendering backend for social video editing, supporting mp4/mov/avi/webm up to 500mb with 30-90 second turnaround. handles token auth, session management, and export workflows via rest apis.

video-to-subtitle automates caption generation and embedding for uploaded video files in multiple languages via a remote gpu backend. supports mp4, mov, avi, webm formats up to 500mb and delivers results in 30-90 seconds with no local software required.

commercial-appraisal-narrative-drafter guides certified general appraisers through uspap 2026-2027 narrative report production for income-producing commercial properties, with six-phase workflow covering engagement intake, market analysis, hbu, three valuation approaches, and certification.

nexus-repository-skill wraps sonatype nexus repository manager 3 rest api operations including repo listing, artifact search, component upload/download, and deletion via a self-contained python cli.

thinking-model-router maps domain and problem type to catalog skills, returning one primary skill or none. routes through mechanism fit via a scored candidate filter with explicit abandon signals.

geoskill-water-purification-mapping applies a simplified Budyko water balance model to compute water yield, retention, and nutrient purification across watersheds, outputting three GeoTIFF rasters. suitable for ecological function assessment and conservation planning.

vercel-ai-gateway wraps the unified inference proxy behind managed maton authentication, exposing 315+ models across 34 providers with per-provider pricing, uptime, and openai/anthropic-compatible endpoints. catalog browsing and inference are both supported.

geoskill-water-table-mapping generates groundwater level and depth-to-water rasters from discrete well observations using idw or simplified kriging with leave-one-out cross-validation. supports terrain constraints and offline processing.

minutes-task provides windows task scheduler integration for sub-hourly polling via python scripts, addressing a genuine workbuddy limitation where byminute rules are ignored. execution flow is concrete but assumes agent automation that may not exist.

ai-life-coach uses socratic dialogue and design thinking frameworks to guide users through self-awareness, goal clarification, and action planning. includes crisis safety protocol for suicidal ideation signals and maintains coaching (not therapy) boundaries throughout.

qa-agent-testing covers ai agent validation across nine dimensions including safety, tool invocation, hallucination detection, and reasoning chains. provides agent-type triage and structured test plan generation with reference to external framework.

video-clip-maker-ai processes raw video footage (mp4, mov, avi, webm up to 500mb) into edited clips via cloud gpu, returning 1080p mp4 files in 30-90 seconds. requires nemo_token authentication and supports iterative editing with session persistence.

modesty-admapix routes ad intelligence queries across multiple api endpoints, detecting user language and executing creative search, app analysis, rankings, and market intelligence via intent classification and multi-step orchestration patterns.