a2a

name	a2a
description	Spawn a team of agentic-CLI sessions (claude, opencode, pi, ...) that talk to each other as peers via a shared SQLite message bus at ~/.a2a/{project}/database.db. No central chain of command — each agent decides who to message. Use when the user wants multiple AI sessions to collaborate, debate, or divide work without a fixed orchestrator.
trigger	/a2a

a2a — Agent-to-Agent Messaging

This doc is a stub. The canonical skill document lives at .agents/skills/a2a/SKILL.md. Ownership and doc maintenance rules are in AGENTS.md. Shell completions are documented in completion/AGENTS.md.

Quick reference

a2a init                          # create project database
a2a register alice --role dev     # register an agent
a2a send bob "hello" --from alice # direct message
a2a send all "broadcast" --from a # broadcast
a2a recv --as bob --wait 30       # receive messages (blocks up to 30s)
a2a peek --json                   # observer view
a2a search <query> --json         # search messages
a2a thread <id>                   # view thread
a2a stats --json                  # bus statistics
a2a status done --as bob          # mark agent done
a2a clear --yes                   # wipe database
a2a-spawn                         # per-CLI launcher for spawning agents

Flags: --project NAME (overrides $A2A_PROJECT > basename of cwd)

Where to go next

• .agents/skills/a2a/SKILL.md — Spawn protocol, kit prompt, full skill spec
• AGENTS.md — Doc ownership table, adding/removing files
• ../completion/AGENTS.md — Shell completion setup
• ../examples/AGENTS.md — Agent pattern examples
• ../README.md — Project overview, install, CLI cheatsheet