Skills for your AI assistant
Add a skill and Claude Code, Codex, Cursor or a similar assistant learns one more job.
Agent frameworks
64 skillsFrameworks, SDKs and harness pieces for agents: tool loops, context management, skill installers.
- eve8k installsBuilds and debugs eve backend AI agents, covering instructions, skills, tools, channels, sandboxes, subagents and evals.
- mcp-developer4.6k installsBuilds and debugs MCP servers and clients that connect AI systems to external tools and data.
- loop-engineering169 installsProvides reference patterns for designing and debugging agent loops with clear stopping conditions.
- ethos-explore57 installsExplores your codebase and asks questions to capture what an AI agent is meant to do.
- ethos-write55 installsWrites and validates an agent's ethos.md file describing its purpose, boundaries and success criteria.
- skillshare49 installsConfigures and runs Skillshare to install, sync and audit skills, agents, plugins and MCP settings across your AI tools.
- ag-ui-a2ui-integration35 installsAdds A2UI rendering to an AG-UI app or framework and wires the middleware to a compatible renderer.
- attach-image31 installsLoads an image file into the model's context so it can see screenshots, diagrams, charts, or photos.
- agent-message30 installsSends direct messages to a parent, sibling, or child agent through the daemon.
- ai-tools-setup10 installsInstalls, connects and repairs Claude Desktop and MCP servers, fixing config files when tools don't show up.
- agui-cross-sdk-parity9 installsPorts features from the TypeScript or Python AG-UI SDK into .NET and checks the wire format matches.
- manage8 installsHandles the lifecycle of custom skills: versions, drafts, publishing, rollback, scopes and approvals.
- triage2 installsBreaks a complex task into stages and picks the right skill for the next step.
- ops-automation-agent1 installsGuides building agents that run scheduled workflows, sync data, send notifications, and automate internal operations.
- agent-evolutionHelps your agent learn from its own run history and improve itself through experience, prompts, code or model updates.
- writing-rig-pluginsWrites, changes or debugs Rig plugins that add slash commands, tools, panels or other behaviour.
- agent-briefWrites clear task prompts for other agents, stating the goal, limits and a checkable definition of done.
- skill-creatorTurns your description of a routine or automation into a new draft skill.
- skill-creatorSaves a workflow or set of instructions as a reusable skill for future tasks.
- agent-budgetsSets an agent's monthly spending budget and limits so mistakes stay cheap and busy months don't halt it.
- agent-capabilityAdds or changes what an agent can do, such as a new tool, knowledge source, web search or approval gate.
- pineSets up and runs event-driven agents that react to messages, files, webhooks, devices and schedules.
- shadcnAdds, searches, fixes and styles shadcn/ui components and builds design systems in your project.
- evidence_synthesisCombines findings from multiple sources into one clear, organized summary of the evidence.
- research_queriesTurns a research question into effective search queries to find relevant sources.
- dsh-crewSplits a large job into roles and a team structure, and sets up sub-agents on the phone to carry it out.
- dsh-reviewHas two or more models from different vendors review the same output independently, then merges agreement and disagreement.
- signet-registryFinds, inspects, verifies, and loads skills and agents from the Signet registry before relying on them.
- project-agentStarts or resumes a persistent project agent that keeps project memory and delegates tasks to other agents.
- ethos-orchestrate54 installsRoutes requests about an agent's ETHOS.md to the right explore or write step.
- goal30 installsReads goal status and budget, starts a goal on request, and marks it complete when done.
- agent-observe29 installsLists an agent's parent, siblings, and children and shows their status and recent messages.
- compact29 installsChecks context usage and summarizes the conversation so the agent can keep working when context fills up.
- rlm-heartbeat29 installsStarts, schedules, and manages agent-owned heartbeats from the Python REPL.
- ambient-healthcare-agent-with-nemotron-voice-agent29 installsCustomizes the Nemotron Voice Agent example into a healthcare intake or tool-calling voice assistant without a separate backend.
- skill-creator28 installsCreates, validates, and installs new markdown or Python-backed skills for the agent.
- mcp22 installsFinds and calls tools on connected MCP services like Notion, Linear, and Slack from Python.
- agui-dojo9 installsRuns the AG-UI Dojo demo viewer locally and registers the .NET SDK as a dojo integration.
- agui-dotnet-agents-sync8 installsUpdates the .NET SDK's AGENTS.md and architecture docs after code structure changes so they stay accurate.
- agui-dotnet-client-tools8 installsLets an AG-UI agent call C# tools that run in the client app and return results to continue the run.
- agui-dotnet-code-review8 installsReviews C#/.NET changes to the AG-UI SDK against its serialization, packaging and style conventions.
- agui-dotnet-cross-language-tests8 installsWrites interop tests that check the .NET SDK exchanges events correctly with the TypeScript SDK.
- agui-dotnet-feature-workflow8 installsLays out the step-by-step workflow and definition of done for implementing a feature in the .NET SDK.
- agui-dotnet-human-in-the-loop8 installsPauses an agent run for human approval or input, then resumes it with the .NET SDK.
- agent-state-barAdds a status bar to your agent so it tracks time, tool calls, TODOs and errors instead of looping or drifting.
- async-event-agentBuilds agents that react to emails, webhooks, timers and alerts, and handle interruptions while tools are running.
- context-compressionShrinks an agent's context when tool results overflow the window, using layered compression strategies and sub-agents.
- context-engineeringHelps write agent system prompts, structure context and author on-demand skills, including defenses against prompt injection.
- error-recoveryHardens an agent against failures like loops, dropped streams, rate limits and context overflow, with retries and circuit breakers.
- loop-engineeringFixes agent loops that stop too early, fake success or never end, using validators and a proposer-reviewer setup.
- mcp-skill-hubHelps connect third-party capabilities to an agent, choosing between MCP and a skill hub and reviewing token cost and security.
- multi-agent-designHelps decide between one agent and many, then design the topology, communication and failure handling for multi-agent systems.
- tool-designHelps you design and describe tools for agents, including parameters and safety checks.
- tool-discoveryLets the agent find and load tools on demand instead of loading them all upfront.
- attach-videoLoads a video clip into the model's context so it can watch screen recordings or short camera clips.
- factoryRuns, watches, stops, and resumes state-machine workflows made of spawned child agents.
- agent-designTurns a request like "I need an agent for X" into a design covering its job, permissions, knowledge and approvers.
- agent-instructionsWrites an agent's instructions so it does one job well, covering role, scope, refusals, sources and tone.
- agent-specChanges the agent specification format, such as adding, renaming or validating fields, while keeping existing agents working.
- platform-contextDecides what belongs in AgenticOS context versus instructions, skills, memory, knowledge, projects and tasks.