Planning.

Planning generates long-horizon plans, maintains goal state, and resolves dependency cycles. Before any high-impact action commits, Planning produces a deterministic dry-run simulation that the operator reviews as a diff against the target state (patent 15). Only on explicit confirmation does the action commit. This eliminates the class of agent errors where the assistant did the wrong thing irreversibly.

• 01Long-horizon plan synthesis
• 02Pre-commit dry-run simulation
• 03Goal-state diffing for operator confirmation
• 04Dependency cycle resolution

Authoritative external corpora and standards this brain treats as canonical. Every retrieval against these sources is signed into the audit ledger so a regulator can prove which evidence drove which output.

• 01Mickai Patent 12, 14, 15, 42
• 02PDDL Planning Domain Definition Language
• 03STRIPS and hierarchical task networks
• 04AAAI Planning track proceedings
• 05ICAPS conference proceedings
• 06Operations Research planning literature
• 07Apache Airflow and Prefect documentation
• 08Goal-oriented action planning (GOAP) literature
• 09MIT operations-research curriculum

Mickai-native tooling primitives this brain implements internally. Codex for sovereign plain-text graph PKM, Lectern for spaced-repetition memory, Stele for citation-provenance, and domain-native primitives layered on top. No external services in the trust path; data stays on operator-personalised hardware.

• 01Loom (DAG plan scheduler)
• 02Sinew (planning-agent framework)
• 03Codex (plan-state persistence)
• 04Marble (dry-run diff canvas)