AI Agents

Coordinate agents around complex tasks

Multi-agent sessions run through five phases: init, analysis, delegation, iteration, and completion. The coordinator breaks the problem down, dispatches agents with the right tools, and assembles the final output.

• delegate_to_agent tool routes work between agents
• Shared scratchpad and status updates keep execution coherent
• Use for architecture reviews, incident RCA, and large refactors

1. Base agent typesSenior Developer, Code Reviewer, Tech Writer, Product Manager, and fully Custom agents.
2. Agent memoryPersistent memory tools let agents recall conventions, recurring issues, and decisions.
3. MCP serversConnect remote MCP servers over streamable HTTP or SSE and their tools appear namespaced in every agent run.
4. Agent identity & ledgerOpt-inEach agent gets a stable identity, a signed short-lived token, and a ledger of everything it touched.
5. Eval harnessOpt-inReplay any past step against a new prompt or model with write tools stripped, then diff what changed.