Neurogenesis

Create a developmental agent from a digital genome: {agent_name, purpose, initial_nodes (unique, >=1), fitness_metrics (>=1), optional growth_rules / safety_axioms}. Returns agent_id + initial graph summary.

Evolve an agent with one task outcome: {task_id, task_type, success_score in [0,1], optional accuracy/user_satisfaction/ cost_efficiency/safety_score/notes/used_nodes/used_edges}. Success strengthens the used edges, failure weakens them; growth and pruning follow the genome's rules u

Full current state: genome, cognitive graph (nodes/edges with weights and trust), and summary counts.

All developmental agents on this mount, with summary counts.

Routing recommendation: the strongest next cognitive steps from a given node, by learned edge weight and trust.

The append-only developmental ledger: every growth, pruning, and evaluation event with reasons (NG3 — returned verbatim).

Portable state document for an agent (import_state recreates it anywhere — including a self-hosted verdigraph-neurogenesis).

Recreate an agent from an export_state document.

Remove a developmental agent from this mount.

Wu Wei compute routing (NG7): register an execution profile — {id, kind?, quality_score [0,1], cost_per_1k_input_tokens?, cost_per_1k_output_tokens?, latency_ms?, gpu_memory_gb?, max_context_tokens?, local?}. A profile can be a local model, cloud API, rules engine, or cached workflow

Choose the cheapest RELIABLE compute path for a task: {id, task_type, expected_input_tokens, expected_output_tokens, min_quality [0,1], risk?, difficulty?, requires_local?}. Hard contract (NG7): profiles below your min_quality are ineligible regardless of cost — compute savings never

Free read: the routing decision log (per-profile counts, estimated costs, reasons) with the Landauer-floor physics context — honest compute-efficiency accounting in the Intelligence Bound frame.

Return capabilities and input contract.