AI & Agents MCP Servers for Claude Code

Local-first control plane for coding agents: routing, reservations, and guardrails.

Token-budget-aware text handling: count, truncate, and chunk for LLM prompts.

Scan prompts for injection attacks, redact PII, and audit LLM SDK usage from any MCP client

Search the skill-builder registry and auto-install skills for Claude Code, Cursor, and Codex.

Find vetted agent skills for your task — privacy-preserving skill discovery over MCP.

Governed AI workflow assets: staged writes, evals-gated template forging, ontology via MCP.

Let Codex orchestrate external coding agents through their native harnesses.

Multi-agent research MCP server supporting OpenAI, Anthropic, xAI, Google, OpenRouter, Groq, and Cerebras

MCP server providing access to 300+ AI models via OpenRouter

Vantio Gate — Policy Latch dry-run evaluate for agents. No live enforce from MCP.

Ed25519-signed receipts for every AI agent action. Local-first, verifiable, no server required.

Deploy full-stack AI agent apps with database, auth, storage, and Edge AI on one platform.

MCP-compliant agentic control framework with 83+ tools for Claude Code and autonomous agent development

Analyzes codebases with tree-sitter and generates AGENTS.md files for AI coding agents.

Read-only, zero-key MCP server for point-in-time AI model API prices. Open data, CC BY 4.0.

MCP server for markdown CRUD operations on specification documents

plan_exact_design_extraction and extract_exact_design_document are tools to use @arc/design's exact-first pipeline directly in arc-mcp

Stdio MCP proxy: intercepts oversized tool results into a queryable embedded DuckDB.

Peer discovery and messaging for Claude Code instances — same machine or across LAN

MCP server: static worst-case token-budget analysis for LLM-agent workflows (LangGraph/CrewAI/OpenAI-Agents) + signed budget certificates.

Give brains to your small models. MCP server enforcing step-by-step Chain-of-Thought.