/super-agent
Turn one big objective into a finished, verified result by running it through the
super-agent orchestrator — a deterministic, resumable, budget-governed state
machine (scripts/super-agent/).
Usage
/super-agent <objective> e.g. /super-agent build a rate-limited login API with tests
/super-agent resume <runId> continue a crashed or earlier run
What it does
Behaves as the super-agent agent and runs the loop:
kodelythecc super-agent start --goal "— always with a budget." --step-cap kodelythecc super-agent next→ perform the action it names:
blueprint skill, … plan --file .
- dispatch → run each ready step (parallel where independent, each on its own branch), … step --status done --branch .
- integrate → conflict-check with scripts/super-agent/integrate.js, merge in mergeOrder, … integrate (or --conflicts … to stop).
- verify → run /verify, then … verify --passed|--failed .
- Repeat until report.
When to use it
- A feature or change too large for one pass, that splits into dependent steps.
- You want parallel execution of independent steps with correct, conflict-checked integration.
- You want the run to survive a crash and resume.
When NOT to use it
- A single edit, one-file change, or a quick fix — just do it, or use the specific specialist.
- Something with no real internal dependencies — the swarm (
/swarm) or a single agent is lighter.
Guarantees
- Nothing is faked: a step that did not succeed is recorded
failed, and a failed
- The budget cap and integration-conflict stop are honored, never bypassed.
verify --passedrequires the tests/build to actually pass, with evidence.