MCP Servers

Provenance labeling for agent inputs, enforced at the tool-call boundary. Blocks untrusted data from reaching shell, filesystem, and network sinks. Ships with an injection corpus and a published list of known gaps.

AI backend engineering guardrails for NestJS, PostgreSQL, TypeORM, and Prisma — task-aware security/architecture/database context for AI coding agents, plus post-task compliance reporting.

YuanSafe (元安全) — a defensive AI-agent hardening skill: static configuration-facing hardening scan of an agent's own runtime (installed skills, MCP servers, tool descriptions, permissions and data surfaces) across three domains — prompt-injection defense /

CDP-native MCP server that lets an AI agent drive Chrome at full depth — deep network/console read, robust interaction, multi-session, and API-flow → runnable-code extraction

The Local-First Spatial Audit & Surgical Revert Layer for Autonomous AI Agents

OpenMetadata context provider for akko-mcp-trino: descriptions, owners, tiers, PII classifications and keys, handed to the agent alongside the columns.

TypeScript LLM router — smart fallbacks, named routes, cost tracking, and caching. Zero runtime dependencies.

A PTY-backed compatibility shim for `claude -p` (Claude Code headless/print mode). Drives the interactive `claude` CLI via node-pty and exposes the same option surface and stream-json contract. Ships as both a library and an `ocp` CLI binary.