Local memory → governed execution → verifiable continuity
I build infrastructure for AI agents that can remember what happened, know what they were authorized to do, prove what they actually did, and show when their authority ended — with evidence that stays local, portable, and independently verifiable.
whitemagic local-first memory & session continuity (MCP, Rust, MIT)
│
▼
continuity-receipt portable, offline-verifiable evidence (open spec, Apache-2.0)
│
▼
mandalaos-gate-lite bounded agent execution that emits receipts (review snapshot)
│
▼
whitemagic-plugins client integrations & adapters
Each repository stands on its own — WhiteMagic does not require MandalaOS, and Continuity Receipt does not require WhiteMagic.
| Want to… | Start at |
|---|---|
| Use it — memory and continuity for your coding agent | whitemagic |
| Review a protocol — spec, vectors, two independent verifiers | continuity-receipt |
| Attack the architecture — containment review snapshot | mandalaos-gate-lite |
Working notes and docs: whitemagic.dev
