Multi-agent task routing: how Organ's agents decide who does what

Description and use

Organ’s multi‑agent router determines work type, owner, and timing using SQL evidence, a routing agent, a review agent, and code invariants, eliminating tool bypass, missing reviews, and duplicates, raising completion to 86.3 %.

Problem

Agents previously chose workflow types and tools, leading to routing around typed tools, bypassing reviews, lost tasks, and duplicate work.

Outcome

The router handled 1,235 requests, completing 909 runs, denying or combining 155, and failing 169. Completion rate 86.3%. Model token cost $708.50 across 1,113 runs, averaging at least $0.77 per started run. The system now records failure causes, separates dedup keys, and tracks denial rates.

Workflow

• Gather evidence in code (SQL)
• Ask routing agent for shape, owner, timing
• Review decision against policy in a separate turn
• Apply invariants in code
• Dispatch or deny with a way forward

Evidence

Agents routed around the typed tools. A deprecated catch‑all tool, labelled "do not use" in its own description, took 457 calls. All five typed tools together took 232.

Source

The type an agent picked could skip review. 32 of 58 OPS generic-task runs, and 18 of 41 from my own marketing department, were really code changes. Generic tasks have no review or validation phase, so that code shipped without one.

Source

The router also doesn't start its own container. Starting one averaged 775 seconds (p90 1,586) when we designed this, so the router runs inside the requesting agent's container, which already exists.

Source

The main problem now isn't that routing breaks. It's that we can't see why it breaks.

Source