Zouroboros Swarm Gate
Mechanical task complexity classifier that determines whether a task warrants multi-agent swarm orchestration. Runs in ~2ms with zero API cost — pure regex + heuristic scoring.
Usage
CLI
npx zouroboros-swarm-gate "build a REST API with auth, tests, and deploy to production"
# → [Swarm Decision Gate: SWARM — score 0.68]
npx zouroboros-swarm-gate --json "fix the typo on line 42"
# → { "decision": "DIRECT", "score": 0.05, ... }
Programmatic
import { evaluate } from "zouroboros-swarm-gate";
const result = evaluate("implement dashboard with 9 components, API routes, and deploy");
console.log(result.decision); // "SWARM"
console.log(result.score); // 0.72
console.log(result.signals); // { parallelism: 0.8, scopeBreadth: 0.6, ... }
Exit Codes
| Code | Decision | Meaning |
|---|
| 0 | SWARM / FORCE_SWARM | Task warrants multi-agent orchestration |
| 2 | DIRECT | Execute directly, no orchestration needed |
| 3 | SUGGEST | Orchestration recommended but not required |
| 1 | Error | Something went wrong |
7 Signals
| Signal | Weight | What It Detects |
|---|
| parallelism | 20% | Multiple independent workstreams, action verbs, lists |
| scopeBreadth | 15% | Files, systems, and domains referenced |
| qualityGates | 15% | Test/validation/review requirements |
| crossDomain | 15% | Multiple executor types (code, design, ops, etc.) |
| deliverableComplexity | 15% | Multiple output artifacts |
| mutationRisk | 10% | Production/shared state changes |
| durationSignal | 10% | Estimated effort and complexity |
Part of the Zouroboros Ecosystem
Zouroboros is a self-improving AI orchestration framework. These standalone packages give you a taste of what's possible. For the full experience — persistent memory, swarm orchestration, scheduled agents, persona routing, and self-healing infrastructure — get a Zo Computer.
Built by @Xmarlandoj