Parallel dispatch

When work splits cleanly into independent slices — frontend, backend, tests, docs — you want one agent on each lane, working in parallel, not three sessions you’re alt-tabbing between. brainclaw dispatches the work, spawns the target agents (Claude Code, Codex, Copilot, …), gives each its own Git worktree so they don’t collide, and tracks every spawn / accept / progress / completed event in the audit log.

# Single dispatch
bclaw_coordinate intent=assign \
  targetAgents=[codex] \
  scope="src/api/" \
  task="Wire the new /v2/auth endpoint"

# Parallel across the active sequence's lanes
bclaw_dispatch intent=execute
# → spawns one agent per open lane, each with its own claim + worktree
# (pass lanes=[…] to restrict; the active sequence is implicit)

Auto-spawn is on by default for same-machine targets; cross-project dispatch (project=<name>) is inbox-only — the target agent picks up the brief async via its own bclaw_work. brainclaw doctor --dispatch surfaces stale spawns and orphan agent_runs.

See parallel feature work for the full pattern in production.