Multi-agent review
Review and merge work from multiple AI coding agents
Keep parallel agent changes understandable when several branches or worktrees converge into one codebase.
The problem
Five clean diffs can still encode five incompatible assumptions.
With Concord
01
Trace
Review each change with its original task and decisions.
02
Sequence
Know which changes depend on others before merging.
03
Close
Mark work ready, reviewed, and handed back into the shared state.
Result
Review becomes integration of known work rather than archaeology across agent sessions.
Common questions
- What goes in a review packet?
- The changes, tests run, risks and decisions recorded by finish_work. Concord writes them to REVIEW_PACKET.md, and concord review-packet prints one for a task.
Further reading
Give your agents shared work-state.
npm install -g @concord-ai/concord-mcp