Packages

Unified hook adapter for AI CLI tools (Claude Code, Cursor, Codex, Windsurf, Kiro, OpenCode, Gemini)

Framework-neutral agent control contract (AGENT-HOOKS-0.1). Canonical Rust core; other SDKs bind to this crate. Cooperative contract, not a security boundary: the host is fully trusted (see SECURITY.md).

Agent Identity Protocol - cryptographic identity and authentication for AI agents

Core identity primitives for the Agent Identity Protocol

Handshake protocol for the Agent Identity Protocol

Agent identity, key management, and trust verification

Structured image diff with JSON output for agent workflows

Agent inbox protocol abstracts queryable data as email.

MCP server for inspecting Claude Code subagent execution logs

AIP — Agent Intent Protocol. Declare an intent, route to the optimal AI provider, settle per-call over x402.

A Rust SDK for building AI agents with multi-provider LLM support

Procedural macros for the agent-io SDK

Claude Code hooks for jj-colocated repos — git guard, workspace creation, and cleanup

WorktreeRemove hook — forgets jj workspaces and cleans up directories

PreToolUse hook — blocks destructive git commands in jj-colocated repos with jj alternatives

WorktreeCreate hook — creates jj workspaces for agent isolation

Kanban CLI for multiple concurrent LLM agents to coordinate on tasks, backed by SQLite

Minimal Agent orchestration kernel for multi-agent discussion

Toolkit for CLI tools integrating with AI agent loops

Simple keyword-indexed knowledge base for LLM agents

Draft platform-native release announcements (HN, Reddit, X, Mastodon, LinkedIn) from CHANGELOG + README (Rust port of @p-vbordei/agent-launch)

Monitor Claude, Codex, and OpenCode Go usage limits from the terminal

A batteries-included Rust library for building agent workflows

Cooperative execution lock to prevent concurrent agent runs on the same key

A sandbox for running coding agents with restricted filesystem access.

A terminal-native music CLI replacing lx-music-desktop, powered by Agentic intelligence.

Agent manifest and capability descriptor — define, validate, and compose agent specs

Convention and validation for committed AI agent memories

Simple key-value memory store for AI agents — persist facts across turns

High-level pub/sub + request-reply API for agent-mesh

Command-line interface for agent-mesh.

Core types for agent-mesh: identity, messages, ACL, Noise handshake

LAN discovery for agent-mesh via mDNS

Cryptographic identity types and signed envelopes for agent-mesh.

Agent Card registry with capability search for agent-mesh

WebSocket relay server for agent-mesh

Client/agent SDK for agent-mesh with E2E encryption

All-in-one agent-mesh server (registry + relay)

Authenticated QUIC transport for agent-mesh (built on iroh).

WASM/PWA client for agent-mesh — cross-platform WebSocket connectivity

CLI for agent-mesh: keygen, register, discover, request, revoke

Local daemon for agent-mesh relay connectivity

Sliding window of recent LLM conversation turns with paired-protection: never drop a tool_use without its tool_result sibling. Zero deps (serde_json only).

CLI messenger for agent-to-agent encrypted communication over Nostr

agent-metamorphosis

AgentMonitorTUI — passive, lazydocker-style TUI for monitoring Claude Code sessions

agent-motif

Messages used for communication between agents.

A fast, terminal-first multiplexer for managing multiple Claude Code conversations across local and remote hosts.

A CLI tool for managing agent notifications