Open source · Prompt & Agent Workflows

Subagent Lane Planner

Split bounded work into parallel agent lanes with clear ownership and outputs.

v0.1.0 · Node.js 22+ · MIT

Browse the public repository · View releases

Splitting work across several agents goes wrong in two places, and both of them look fine on paper.

This walkthrough uses the tool's public README and checked-in example files. Run the command from a repository checkout with Node.js 22+.

Run the checked-in example

# A plan that can be dispatched: three waves, one blocked lane. Exits 0.
node bin/subagent-lane-planner.mjs \
  --plan examples/plans/feature-split.json --root examples/sample-repo

# The same run, writing the lane packets outside the repository.
node bin/subagent-lane-planner.mjs \
  --plan examples/plans/feature-split.json --root examples/sample-repo \
  --packet-out build/example-packet.json

# A plan that cannot be dispatched as written. Exits 1.
node bin/subagent-lane-planner.mjs \
  --plan examples/plans/unbuildable.json --root examples/sample-repo

# A prerequisite nobody has checked. Exits 2, incomplete, no plan.
node bin/subagent-lane-planner.mjs \
  --plan examples/plans/unchecked-prerequisite.json --root examples/sample-repo

Read the result

The two shapes of exit 2 are deliberate. A configuration error means the run never had a subject, so there is nothing to report about. An input that could not be read means the run had a subject and failed to obtain evidence about it, and a consumer needs the report to know which input was not read.

Scope

Limits may also be set in a --config file ({ "schemaVersion": "1", "limits": { ... } }); a flag overrides the file, and an unknown key in either is refused.