Context.

Context is the assembler. For every cooperative call it synthesises the working context for a downstream brain by drawing from Long-Term Memory, Retrieval, Embeddings, and any active artefacts. It enforces clearance boundaries, redacts content the recipient is not authorised to see, and signs the assembled context so the receiving brain can verify it.

• 01Context assembly from memory, retrieval, and artefacts
• 02Clearance-boundary redaction at assembly time
• 03Signing the assembled context for downstream verification
• 04Token-budget management

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 05, 18, 20
• 02Anthropic Constitutional AI paper
• 03Context-compression academic literature
• 04Token-budget management literature
• 05LongBench and RULER long-context evaluations
• 06Information-theoretic compression bounds
• 07Needle-in-a-haystack benchmarks
• 08Anthropic context-window engineering notes
• 09Open-source RAG-context assembly libraries

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.

• 01Codex (context-source plain-text)
• 02Tablet (block-level context references)
• 03Stele (source-provenance graph)
• 04Aegis (clearance-boundary redaction)
• 05Cipher (signed assembled context)