Search

Type-safe authorization for TypeScript with no classes and no hidden state. Rules are plain JSON, cross the server/client boundary as-is, and compile to SQL. The guard checks a server action, an HTTP handler or an agent tool call. Zero dependencies.

Task-specific React apps for DeepSeek Harness with state carried into the next Agent turn

Birta CLI and local MCP server

Smithers command-line interface, MCP server, and local workflow tools

MCP server that connects AI assistants to Chrome DevTools Protocol for runtime debugging - set breakpoints, inspect variables, monitor network traffic, and automate browser interactions

MCP server that lets an LLM analyze, judge, and choose chess moves (Stockfish + Maia3 + Lichess)

Semantic search over RSS/news feeds for AI agents — a single-file MCP server, no API key, no cloud.

VibeKit MCP client Tool

Model Context Protocol (MCP) JSON-RPC safety clearance middleware for AI agents

Anthropic Claude tool_use deterministic invariant safety guard

stdio MCP server for z-library search & download — bring your own account

Deterministic, local-only developer primitives for AI agents. No network, no filesystem, no telemetry.

MCP server for CODASMS — receive one-time SMS/OTP verification codes (get a number, wait for the code, release) across 700+ services and 180+ countries via the CODASMS Reseller API.

MCP server for App Store & Google Play app data (search, details, IAP prices, reviews, top charts, estimates) — powered by the App Intel API on RapidAPI. Works with Claude Desktop, Claude Code, Cursor, Windsurf.

https://github.com/CSAO-ORG/csoai-static-deploy2

Model Context Protocol server for Railyard — read and write access to Railyard datacenter projects and organisations over stdio, authenticated with a personal access token.

MCP server for DMARCReport API - enables AI assistants to manage DMARC configurations

Relay channel for Claude Code. Bridges a Relay agent conversation into a Claude Code session over the experimental claude/channel MCP contract.

MCP server for Lightning FM's Nostr-native music catalog — search tracks, look up artists, and check what's live now. Wraps the same relay + Blossom surfaces the platform itself uses; no separate agent API.

An accessibility auditor built on a 3-agent pipeline (Planning, Execution, Validation) that scans a page, grounds fixes in WCAG/ARIA guidance via local RAG, and validates fixes by simulating them in a cloned DOM before reporting them as confirmed.