23,553 tagged Building agents, measured the same way as everything else here.
Browse within: agentic-ai 233LLM 217agent 200agents 151skills 145agentic-workflow 110agent-orchestration 86openai 86antigravity 79agentic-framework 73cli 73ai-skills 71Autonomous Agents 67agent-framework 63
rafsilva85/manus-credit-optimizer
Skill Claude CodeCodex
Automatically optimize AI agent credit usage by routing tasks to the most cost-efficient execution path. Use when you want to reduce AI API costs by 30-75% without quality loss, classify task complexity before execution, route simple tasks to free or low-cost models, split complex tasks into optimized sub-tasks, or…
Skill Claude CodeCodex
Persistent context and inter-agent messaging across sessions and platforms. Activate when the user asks about resuming previous work, continuing where they left off, sharing context between agents, multi-agent coordination, remembering across sessions, cross-platform context sync, agent communication, or mailbox…
Skill Claude Code
Run a Threadwork orchestration pattern (research-then-write, parallel-review, code-review-fix) over your yaml-defined roles. Use when the user asks to chain multiple roles, review work across reviewers, or iterate on code with a feedback loop.
Skill Claude CodeCodex
Delegate engineering work to DeepSeek subagents that read and write the current project directly. Seven tools (agent, orchestrate, agentcontrol, memory, review, generate, analyze) over a self-filling markdown memory — one subagent for a bounded task, a dependency graph of them for a plan, and facts about the codebase…
orchestratemcp/OrchestrateKIT-MCP
Skill Claude CodeCodex
Plan safer AI-agent workflows with OrchestrateMCP. Use when the user wants to design, scope, safety-review, or hand off an AI automation, agent, MCP workflow, or multi-step tool workflow.
AlgoChains/algochains-mcp-server
Skill
MCPorter toolkit for AlgoChains — CLI, SDK, multi-MCP composition.
Skill Claude CodeCodex
Develop, test, and manage shell command bundles as standard MCP tools via YAML + Bash. Use when creating a new command bundle, adding or modifying individual tools, debugging YAML spec or script issues, running quality gates, or deploying the server.
Skill Codex
Use a paired AgentBridge ChatGPT webpage for bounded independent review, planning, or adversarial critique when a local coding agent needs a user-authorized second perspective. Do not use it for secrets, direct execution, or routine local implementation.
Skill Claude CodeCodex
Demonstrates the expected Claude-style skill layout for skills-as-mcp.
Skill Claude CodeCodex
A Chinese-language skill for building and troubleshooting MCP servers that work with the StepFun AI desktop assistant.
Skill Claude CodeCodex
Structured 10-domain identity interview that maps operator values, preferences, and boundaries into HomarUS identity files (soul.md, user.md). Onboarding experience for new agents.
Skill Claude CodeCodex
Set up and optimize OpenClaw agent workspaces using the clawspace CLI.
bakalagoin/codinfy-agent-monitor
Skill Cursor
Monitor AI agents, context, limits, workflow, Git, tests, build, secrets, and model economy.
Skill Claude CodeCodex
Introspect the user's local AI agent infrastructure in real time. Use proactively: before starting tasks, before scheduling cron/timer work, when stuck, after long tasks, or when claiming something works. Exposes tools: vitalssummary, shadowlist, shadowstale, burnoutsummary, burnoutstucksessions.
Skill Claude CodeCodex
Instructions for using an agent that guides an enterprise AI project through discovery, assessment, architecture, prototype, and handoff stages. The process produces files and other structured deliverables, with analysis supplied by the host agent.
Skill Claude CodeCodex
Delegate bounded processing tasks to the local cluster (0 cloud tokens) via mcplocaltokens tools. Use BEFORE generating boilerplate, docstrings, mechanical tests, file summaries, first-pass code reviews, format conversions or commit messages.
Skill Claude CodeCodex
Route a domain question to the right expert(s) on the team and return a cited answer. The team and what each expert knows.
Skill Claude CodeCodex
Use when the user asks you to team up with another agent session (same or different agent), send/receive messages to/from other Claude/Codex/Cursor/Antigravity sessions in the same project workspace or a shared room, act as reviewer for another agent, or join a multi-agent team. Triggers on phrases like "hỏi session…
AMATSUGI-120/plugin-health-auditor
Skill Codex
Audit Codex plugins, Agent Skills, hooks, scripts, and MCP configuration for always-on triggers, context overhead, recursive invocation, subagent amplification, instruction conflicts, risky commands, excessive capabilities, exposed secrets, and malformed manifests. Use only when the user explicitly asks to audit…
Skill Claude CodeCodex
Prove Before Act proof and accountability integration for AI agents. REST API, MCP, and x402. Check live endpoints for availability, pricing, and proof status.
Skill Claude Code
Decompose a task into a DAG of subtasks executed by parallel CLI agents. Use when the user asks to cascade a task, build an agent graph, or fan work out across parallel agents. Trigger — /cascade.
Skill Claude CodeCodex
Trigger: agent orchestrator, registeragent, listagents, agentsend, agentread, agentstatus, talk to opencode, multi-agent MCP. Register this running agent instance with the shared Agent Orchestrator MCP hub and message another registered agent, bidirectional and non-blocking.
JacobLinCool/make-chatgpt-site
Skill Codex
Discover, inspect, create, edit, version, and deploy ChatGPT Sites associated with a local project or repository. Use when an agent needs to find a repository's .openai/hosting.json, identify its linked Site, inspect existing versions or deployment state, create a new Site, modify the local website, save a new…
lenhonbp/codex-memory-intelligence
Skill Claude CodeCodex
Apply the CMI Agent Operating System contract for substantive work: orient with explicit goal, scope and authorization boundaries; observe before editing; capture typed evidence; diagnose before symptom fixes; prioritize explicitly; implement the smallest coherent change; verify the original journey or contract…
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: