deai-marketplace connects agents to a decentralized asset auction platform on base, offering shell scripts for registration, browsing auctions, bidding, buy-it-now purchases, and settlement across multiple token standards with atomic on-chain execution.

service-watchdog monitors http endpoints, tcp ports, ssl certificates, and dns resolution for self-hosted infrastructure, emitting structured alerts and trend data via configurable bash commands.

geo-tag-photos recovers gps metadata for jpg images by identifying landmarks via vision and geocoding them against openstreetmap. dry-run by default, requires explicit backup and write flags to modify files.

skill-release-plus publishes skill packages to multiple hubs (clawhub, tencent cloud, github) in one command, with pluggable adapters for custom destinations, single-pass signing, and configurable exclude rules.

regime-gated donchian short that fires only during triple-confirmed strong-bear conditions (ema separation, adx, 30-bar momentum). validated 6 trades, 100% win rate, zero drawdown on btc over 162 days; produces zero signals in chop by design.

allowance handles purchase approval and virtual card issuance for agent-driven shopping, with local browser checkout and phone-based verification. covers immediate and delayed purchases across merchants.

idea-to-design enforces collaborative design before implementation through a structured 9-step process including context exploration, clarifying questions, approach comparison, and written spec validation. requires user approval at multiple gates.

math-help maps five computational layers (sympy, z3, scratchpad, tutor, lean4) with concrete command examples and decision trees. covers symbolic algebra, constraint solving, reasoning verification, learning, and formal proofs.

mobile-design enforces platform-aware constraints before implementation, mandating reference materials and blocking common performance, touch, and security anti-patterns. it covers feasibility assessment, mandatory checkpoint gates, and release readiness validation across ios and android.

realbrowser lets ai agents drive real chrome sessions for authorized tasks like qa automation, synthetic monitoring, and data collection on sites where you have explicit permission. operates in self, marketplace, or earn modes with transparent billing and privacy boundaries.

fragment-thoughts-organizer-en catches scattered thoughts, half-sentences, and snippets throughout the day and files them chronologically without rewriting. it surfaces latent connections across semantic, temporal, emotional, people, and scene threads without interpreting for the user.

xmtp-agents bridges any agent runtime (openclaw, claude code, langchain, custom python) to xmtp messaging with owner/public role separation. incoming messages stream through your agent backend for full-capability responses while public users operate in restricted conversation-only mode via customizable prompt files.

ocas-taste extracts consumption signals from email and calendar, enriches venue metadata via maps and search, weights signals by recency and frequency, and generates evidence-backed recommendations that cite specific prior behavior and respect dietary constraints.

nexus-ciba-wechat generates shareable html and png cards from the iciba daily sentence api, supporting multiple templates (vertical moments card, horizontal postcard) with qr-encoded audio links and fallback image generation via playwright.

postzee generates ai images and videos, optimizes prompts, and distributes content across 30+ social platforms. covers image-to-image and image-to-video workflows, heygen avatar creation, and multi-channel scheduling.

mergeguard provides structured pre-merge review for ai-generated code across github prs, local diffs, and pasted patches. it enforces a strict decision gate (merge/fix first/reject/blocked) and covers correctness, security, scope, tests, and risk calibration with concrete output templates.

frontend-slides builds zero-dependency, browser-native HTML presentations with animation, viewport constraints, and deliberate design. covers new decks, ppt conversion, and enhancement of existing slides with style discovery workflow and cross-platform delivery.

openai-newapi-calc-token provides formulas and decision trees for computing token costs across openai-raw, newapi-quota, and custom-multiplier billing modes, with explicit clamping rules and common error patterns.

ui-audit provides a prioritized checklist workflow for web interface quality review, covering accessibility, interaction, forms, typography, and layout across changed surfaces with structured rule categories and concrete fix guidance.

skill-dreaming-extractor processes daily agent session logs to extract structured facts (decisions, problems solved, config changes) into a persistent learned.md file, with token budgeting and confidence scoring to build a searchable knowledge base.

repomix-rs-explorer packs local or remote codebases into single ai-friendly files for high-level analysis. covers structure discovery, pattern matching, and metrics collection across unfamiliar repositories without requiring targeted edits.

cruit-recruiter guides users through structured sourcing and outreach on the cruit platform, with explicit permission gates before spending credits or revealing contact info. covers authentication, search filtering, candidate matching, and direct messaging.

ma-diligence-issues-list transforms raw data-room artifacts into evidence-anchored, severity-rated issues tables with top-10 summaries, management q&a, and structured r&w recommendations for buy-side deal teams.

local-inference-context manages kv cache saturation on self-hosted llm backends (llama.cpp, ollama) by calibrating effective context budgets, monitoring vram headroom, and defining threshold-based recovery actions before 503 errors occur.

interview-synthesis guides researchers through extracting observations, clustering themes, selecting quotes, and producing stakeholder-ready synthesis reports from qualitative data. includes routing for interviews, usability tests, focus groups, and surveys.

mjzj-skillhub surfaces a cross-border ecommerce skill marketplace with search, filtering by tags and price tier, and a managed submission workflow for new skills. integrates four apis covering public queries, label taxonomy, file uploads, and skill application review.

git-worktree-manager standardizes parallel branch isolation by automating worktree creation, deterministic port allocation, environment sync, and safe cleanup. designed for multi-agent workflows needing concurrent feature work without port or state conflicts.

wshobson-architecture-patterns covers clean architecture, hexagonal architecture, and domain-driven design tactical patterns with concrete code examples, layer dependency rules, and anti-patterns to avoid in practice.

aevo-trading-skill connects mcp clients to a 45-tool derivatives trading platform for perpetual futures, spot, and options. it emphasizes capital preservation through mandatory pre-trade risk checks, confirmation flows, and multi-leg strategy support on a decentralized exchange.

adversarial-review spawns parallel claude-opus reviewers with distinct critical lenses on substantial documents, collects structured redlines, and guides positioning to produce a revised v2. self-triggers on complexity thresholds or explicit user cues.

shipguard enforces a six-gate workflow for ai-assisted development: requirement intake with test cases, impact analysis, typed implementation, feature qa, regression testing, and lessons capture. prevents silent breakage and scope creep in shared codebases.

secondhand-furniture-listing-check generates a practical readiness sheet for evaluating used furniture listings, covering seller questions, photo evidence, fit verification, hygiene and pest warnings, recall checks, and safer pickup logistics without legal advice or valuation claims.

statistical-analysis provides actionable guidance on descriptive statistics, trend detection, outlier handling, and hypothesis testing, with explicit warnings on correlation, multiple comparisons, simpson's paradox, and survivorship bias.

epub-bilingual-converter-skill converts ebook files into bilingual editions through staged extraction, token estimation, translation, and assembly, preserving structure and enabling paragraph-aligned target-language content with summary generation.

google-analytics covers both admin (write) and data (read-only) apis via maton proxy. manages oauth, requires explicit approval for mutations, provides reference endpoints and common report templates.

loseit extracts per-day nutrition data from lose it exports as json, supporting both downloaded zip files and cookie-based fetches. read-only, no dependencies, outputs calories, macros, meal breakdowns, and budget figures.

onenote covers notebook, section, page, and content management via microsoft graph api. supports read and write operations across personal, sharepoint, and group notebooks with explicit confirmation workflows for mutations.

dataify-youtube-profiles wraps the dataify builder api for youtube channel collection by url or keyword, handling token resolution, parameter validation, and task submission with explicit workflow guidance.

authenticate-openant covers key-based and email otp login flows for new and existing agents, with agent registration, identity retrieval, and email binding. prerequisite for all write operations in the openant platform.

sarif-parsing provides comprehensive strategies for extracting, aggregating, and deduplicating static analysis findings using jq, python libraries, and fingerprinting techniques. covers path normalization, schema validation, and ci/cd integration patterns.

linkedin-post-openclaw-browser automates linkedin feed post publishing via browser automation, enforcing a prepare-then-publish workflow with explicit user sign-off before external side effects occur.

rustunnel exposes local http and tcp services through public urls via mcp tools. covers webhook testing, demo sharing, ci/cd preview environments, and database access with automatic lifecycle management and proper cleanup.

context-verifier computes file hashes, detects integrity drift, and generates verifiable context packets for review workflows. prevents agents from editing stale files by confirming checksums before action.

swiftui-view-refactor establishes patterns for decomposing SwiftUI views into small, explicit types with preference for model-view over view-model architecture, stable view trees, and extracted side effects.

review-agent-governance gates ai agent actions on github and external surfaces behind explicit human approval, producing signed receipts for audit. targets pr reviews, issue triage, releases, ci edits, and protected branch pushes.

prometheus-alert-designer covers designing actionable alerting rules, multi-window burn-rate slo alerts, recording rules for performance, and audit patterns for reducing alert fatigue in production prometheus deployments.

supabase covers local-first development with the cli, migrations in version control, row-level security patterns, edge functions, and storage setup. includes deployment workflows and anti-patterns to avoid.