Menu
Create, test, and improve functions at runtime — agents write their own code
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
Session lifecycle management — transitions, state tracking, reactions
Persistent agent memory — store, recall, search across sessions
Plan and execute multi-agent work — decompose features, spawn workers, monitor progress
Scan agent health and auto-recover stale or dead sessions
Security scanning — injection detection, capability checks, audit trails
Multi-agent swarm coordination — create swarms, broadcast tasks, reach consensus
| name | evolve |
| description | Create, test, and improve functions at runtime — agents write their own code |
| toolScope | ["Bash","Read"] |
| effort | extended |
Self-evolving functions:
curl -X POST http://localhost:3111/api/evolve/generate -H 'Content-Type: application/json' -d '{"goal": "<what>", "name": "<name>", "agentId": "claude-code"}'curl -X POST http://localhost:3111/api/evolve/register -H 'Content-Type: application/json' -d '{"functionId": "<id>"}'curl -X POST http://localhost:3111/api/eval/run -H 'Content-Type: application/json' -d '{"functionId": "<id>", "input": {...}, "expected": {...}}'curl -X POST http://localhost:3111/api/feedback/review -H 'Content-Type: application/json' -d '{"functionId": "<id>"}'Functions evolve through: draft -> staging -> production. The feedback loop auto-improves underperforming functions.