Research
Not Human Search
MCP server for discovering other MCP servers. Indexes 8,600+ agent-friendly sites with agentic scoring. Lets Hermes find new tools on its own.
Automation
Agent Desktop
Desktop automation via native OS accessibility trees using the agent-desktop CLI. Use when an AI agent needs to observe, interact with, or automate desktop a...
Automation
Perkoon Transfer
The agent data layer. Perkoon moves files between agents and the physical world — agent to human, agent to agent, agent to pipeline. P2P over WebRTC. Free. E...
General
Stellar Trails
Activates on every task: coding (features, bugs, refactoring, scripts), documents (reports, proposals, DOCX, PDF), charts and visualizations, data processing...
DevOps
FDE
前线部署工程师——梳理企业工作流、识别 AI 节点、构建知识库、安装 sofagent 底座、交付离场。 内置持续优化模式（sustain），自动读 audit 报告趋势生成优化报告。
Productivity
Skill
当你的 Agent 反复偏离目标、任务越做越复杂、刚踩过的坑下次还踩。装上后 Agent 自带行为约束和反思能力。
Business
8004skill
Manage on-chain AI agents via ERC-8004 — register, search, load, update, inspect, transfer, give feedback, check reputation, set wallets, verify identity, send A2A messages, execut…
Automation
autopilot
Use this skill whenever the user wants to ship a side project end-to-end on a deterministic, repeatable agent harness. Also trigger when the user says "/autopilot", "ship my side p…
AI / ML
agent-framework-builder
Use this skill when the user wants to set up, plan, or scaffold an OpenCode agent framework for a new project. Triggers on phrases like "build an agent for this project", "set up a…
Engineering
agent-api-design
Expert in designing APIs, CLIs, and documentation systems for AI agent consumers. Use when building or reviewing interfaces intended to be used by LLM agents rather than humans — A…
Security
symbiont
AI-native agent runtime with typestate-enforced ORGA reasoning loop, Cedar policy authorization, CommunicationPolicyGate for inter-agent governance, ToolClad declarative tool contr…
Automation
canvas-agent-chat-skill
Call a local or remote Canvas Agent `/chat` API and return either the final answer or a structured `question_required` payload. Use this whenever another agent should delegate work…
General
first-party-skill-scout
Use when LawFirm OS needs to discover candidate agent skills, compare multiple public SKILL.md examples, respond to an Exception Lake skill-gap candidate, or draft a safer first-pa…
DevOps
durable-context-maintenance
Keeps repo-local agent context accurate after code and workflow changes. Use when docs/ai, AGENTS.md, runbooks, or context entrypoints have drifted from the codebase, after major f…
Security
auditing-skills
Scans AI agent skill directories for dangerous bash patterns, prompt injection, supply chain risks, and SKILL.md structure violations. Use when reviewing, validating, or security-a…
Engineering
decomp-repo-agent-cloneable
Use when decomposing a multi-workflow project into standalone, agent-cloneable repos. Codifies the 7-step pattern proven on career-ops, aicreator, StockForge, and mission-control. …
Automation
loop-design
Generates executable loop specs by filling the seven B-type loop engineering elements—context, feedback, verification gate, termination, error handling, state, and cost guard—plus …
AI / ML
add-codex
Use Codex (CLI + AppServer) as the full agent provider — planning, tool orchestration, native compaction, MCP tools, session resume — in place of the Claude Agent SDK. ChatGPT subs…
Engineering
ghs-issue-implement
Implements GitHub issues via worktree-based agents and opens a PR. Clones the repo, creates a worktree, runs an agent to build the fix or feature, verifies the result, and submits …
AI / ML
agoragentic
Use when an autonomous agent or builder needs to preview Triptych OS deployment, export a Micro ECF harness packet, execute routed work, resolve agent:// aliases, discover capabili…
Automation
lark-vc-agent
Join or leave ongoing Lark video meetings on behalf of a user and capture live events such as participant arrivals, departures, speech, chat, and screen sharing. Use +meeting-join …
Automation
agent-team-launcher
Selects, starts, or resumes an agent team initiative from .foundry/staging/. Scans configured initiatives, displays status, lets the user choose, then outputs the terminal command …
Productivity
sandeco-loop
Creates a hardened agent loop specification document following the Loop Library pattern (trigger + goal + verification + stop). Captures full definition—use cases, inputs, meta, ch…
General
vertragshaendler-abgrenzung
Determines the distinction between commercial agent and authorized dealer: independent dealer purchases in own name and for own account with no statutory compensation under § 89b H…
Automation
bailout
Forces continuation when an agent halts mid-task, defers for direction, deletes work without a plan, or collapses after setbacks. Interrupts the pattern, surfaces hidden issues, na…
Automation
bridge-agent-core
Core skill for all Bridge IDE agents. This skill should be used for bridge communication, task management, agent coordination, and understanding the Bridge IDE platform. Covers bri…
Testing
verify-agent-output
Validates deliverables and builds evidence during agent handoffs, including pre-execution expectations, structured proof collection, post-execution validation against external anch…
Productivity
agent-recall
Use Agent Recall to find prior coding-agent conversations across Claude Code, Codex, and OpenCode. Use proactively when context may exist in another chat, session, or agent; when t…
Automation
linggen
Durable cross-host memory and browser control via one local MCP server. Three-tier memory model stores user identity (core + long-term + episodic) rather than activity logs; browse…
Automation
ralph-lisa-loop
Orchestrates plan-implement loops with expert review via subagents for planning, implementation, and self-review, plus external review. Human steers; a single knob tunes interrupti…
Automation
loop-design-check
Designs and reviews goal-oriented agent loops for failure modes like token waste, verifier gaming, and completing wrong answers. Writes loops with decidable goals and skeletons; au…
Productivity
analyst-modes
Defines five analyst operating modes (Query, Dashboard, Document Compile, Extract, Challenge) with procedures, output formats, and metrics. Includes write-gates for dashboard and s…
Automation
adspirer-agent
Paid-media advertising agent that creates, analyzes, and optimizes campaigns across Google Ads, Meta, TikTok, LinkedIn, Amazon, and ChatGPT Ads. Enforces paused-by-default safety r…
Security
cursor-agent-safety
Audits and hardens agent run-mode policies across compatible coding agents. Analyzes current settings, generates project-specific classifier rules, builds terminal command allowlis…
DevOps
candy-doctor
Diagnose Candy health status across agent setups—dependency integrity, login validation, JOBS_ROOT structure, agent registration consistency, stale logs/csvs, optimizer gate, macOS…
Business
makler-abgrenzung
Clarifies the legal distinction between commercial agents and brokers under German law: ongoing contractual relationship versus one-time mediation, fixed commission versus success-…
AI / ML
self-improvement-loop
Self-improving orchestrator that runs periodic cross-project reviews, identifies root causes, creates improvement issues, and executes them end-to-end. Operates in discovery mode f…
DevOps
agent-readiness
Assess, harden, or scaffold a repository for coding-agent harnesses. Generates AGENTS.md, SKILL.md, and MCP scaffolding plus hooks, CI gates, and sandbox design; ships a research-b…
Automation
prime-the-agent
Fast 30-second session starter that loads the current site, detects the page builder, ingests inline schemas, and enforces the no-raw-HTML rule. Maintains persistent per-site memor…
Automation
dynamic-workflow-mode
Designs task-local harnesses, eval gates, and reusable skill extraction for dynamic workflow mode and adaptive agent harnesses. Use when a task needs a custom loop, evaluator, watc…
General
alleinvertreter-alters-krankheitskuendigung
Assesses sole agent's rights under § 87(2) HGB: entitlement to district commission on direct sales, distinction from exclusive distribution rights, burden of proof for appointing a…
AI / ML
self-improving-agent
Continuous learning system that captures atomic instincts with certainty scores, scopes them per-project or globally, and evolves them into reusable skills, commands, or agents. Ac…
Automation
agent-rules
Captures agent failures from eval runs or telemetry into per-file reflection logs, then promotes recurring patterns into rules or CI gates once they hit the three-entry threshold. …
Research
agentprivacy-theia-derivation
Domain skill for origin-as-impact patterns in agent derivation. Explains how agents arise through separation events, how derived agents carry novel material, and how violent origin…
Design
design-for-agent-users
Designs or reviews software, docs, SDKs, and repos so AI agents can consume them effectively. Covers agent-readable documentation, SDK patterns for tools and structured output, rep…
Automation
execute-agent-workflow
Executes a 7-step workflow for implementation tasks—bug fixes, refactoring, performance work, or new features—when given a task request or issue URL. Manages overall control flow a…
AI / ML
brain-clone
Converts a domain expert's knowledge into a complete agent via multi-phase interview, capturing identity, reasoning patterns, tools, examples, and output formats, then generating p…
Productivity
cursor-rules-authoring
Defines procedures for authoring and reviewing agent rules files: activation modes, glob vs intelligent scoping, content patterns, anti-patterns, and checklists. Use for creating, …
Security
llm-agent-security-principles
Security principles and defense patterns for LLM agents: Security by Absence, Untrusted Content Boundary, and One External Adapter. Apply during design, architecture reviews, capab…
DevOps
buildkite-agent-runtime
Enables Buildkite pipeline steps to annotate output, manage artifacts, share metadata, upload dynamic pipelines, request OIDC tokens, acquire locks, update step attributes, redact …
Automation
session-to-agent
Converts interactive coding session goals and processes into a reusable goal-seeking agent with memory. Extracts goals, constraints, and approaches from transcripts and generates a…
Automation
verdikta-bounties-onboarding
Guides autonomous agents through Verdikta Bounties setup: create a wallet, fund with Base ETH, swap a portion to LINK for fees, optionally sweep excess, and register for the Agent …
Productivity
aif-skill-generator
Generate professional Agent Skills for AI agents. Creates complete skill packages with SKILL.md, references, scripts, and templates. Use when creating new skills, generating custom…
Content
gerador-soul
Generates a complete, structured SOUL.md file for any AI agent or assistant. Use when creating a SOUL.md, defining agent identity, describing the 'soul' of an assistant, or respond…
Testing
check-skill-pair
Audits integrity of a generated skill pair by verifying all four required artifacts—principles doc, build and check SKILL.md files, and routing registration—then flags cross-artifa…
Automation
mm-bridge
Hands off work to another agent instance by generating a ready-to-copy prompt. Auto-injects project context from passport.md and supports standard prompt frameworks. Outputs to a s…
AI / ML
add-function-agent
Creates a single business-function agent (billing, support, recruiting, compliance, CRM, analytics, marketing) backed by a verified-write MCP integration. Includes explicit scope, …
Testing
agentic-a11y
Multi-agent accessibility audit covering web, mobile, desktop, CLI, and games. Dispatches specialists for screen reader, vision, motor, cognitive, and multimedia issues, validates …
Automation
agent-composer
Plans optimal agent dispatch by analyzing current context, asks 1-2 clarifying questions when needed, and auto-records results after each wave. Triggered by "compose agents", "whic…
Automation
agent-comm-hub
Python-based multi-agent communication hub with MCP+SSE for real-time messaging, task scheduling, shared memory, and evolution. Includes 53 MCP tools, four-tier permissions, and ze…
Showing the top 60 of 1,164. See the full list →