The brain: what every coin's agent knows, sees and learns

One shared brain for the pad, kept by a VectorPad main agent, that every coin's agent draws on; plus a daily reflection per coin. Code: gateway/mirrorpad_gateway/brain.py, canon: gateway/mirrorpad_gateway/canon/*.md, tests: gateway/tests/test_brain.py.

   canon/*.md ──render(settings)──► canon-* facts ──import_facts (versioned)──► every coin's memory (trusted, guard-proof)
                                                                              ▲
   registry ──pad_digest──► main agent (role "brain") ──► learned.md ──"pad-lessons"─┘   nightly
   registry ──write_brief──► PAD.md in each workspace                                     every 2 min
   day_digest ──► the coin's community agent ──► journal/<date>.md + PLAN.md             daily per coin

Canon: what every agent knows

Nine templated markdown files, rendered from the gateway's settings so they never drift from what is deployed: what VectorPad is, the launch rail and the pool ({{venue}} is MIRRORPAD_LAUNCH_VENUES' first entry; the paragraph for a coin follows the coin's own source, Meteora, Orca or simulated), the fee split, the DAO thresholds, the desks, the builder, safety, Solana basics, and playbooks per role and archetype. Placeholders are {{name}}; an unknown name raises (render), so a typo fails tests rather than shipping. Each file is a fact with a content-hash version; import_facts now remembers key@version, so a changed file is re-imported on the next sync and Mirror retires the old version under the same key (verified with the CLI: same key, status retired v1, active v2). The guard never rolls a canon fact back: imports hold the agents' turn locks, and the ledger rule keeps anything older than a snapshot.

Sync points: adopt() right after the official-links import, and after every rewrite of the lessons. GET /brain reports the canon version and keys; GET /tokens/{mint}/brain lists which versions a coin carries.

The system prompt (persona.py) now says what VectorPad is without naming a venue, describes the coin's own pool from venue_facts(settings, token.source), and points the agent at its canon memories, PAD.md, PLAN.md and journal/.

Brief: what is happening right now

PAD.md in every coin's workspace, rewritten every MIRRORPAD_BRAIN_BRIEF_INTERVAL (120 s) by the gateway (it is a gateway-owned file; agents read it): the rail and coin count, the coin's numbers and rank by influence, its plan, the latest questions in its room with who asked (dev, holder and share, visitor), its open proposals and last decisions, its latest calls, what is moving on the pad, other coins' latest calls and posts, and the first lines of the pad's lessons. The prompt tells the agent to read it before posting or answering about the pad.

Reflection: journal and plan, daily per coin

reflect_loop picks one due coin a minute (last journal older than MIRRORPAD_BRAIN_REFLECT_INTERVAL, 24 h). The gateway writes brain/day.md (the coin's posts, mentions, questions and their themes, DAO, desk, events, numbers) and asks the community agent, unguarded (its own trusted turn, so what it saves with save_memory stays), for exactly:

## Journal
<120 to 200 words>
## Plan
- <action>  (anything needing a vote starts with "propose:")

parse_reflection tolerates a reply without headings (all journal, no plan). The gateway writes journal/<date>.md and PLAN.md (the community role has no write tool by design), fires a reflection activity event, and the next brief quotes the plan. The dev can trigger one: POST /tokens/{mint}/brain/reflect. The site can show journal and plan from GET /tokens/{mint}/brain (no UI yet).

Learning: the pad's lessons, nightly

learn_loop (every MIRRORPAD_BRAIN_LEARN_INTERVAL, 24 h, first run 30 min after boot if there are coins) compiles pad_digest: calls over 7 days by score band and tag with outcomes, posts with the most replies, what holders ask (word themes and samples), proposals by kind and status. The main agent reads it plus the previous lessons and the canon in its workspace (data/tokens/pad/brain/) and writes under 450 words in four sections (what worked, what holders ask, desk patterns, for every agent). The gateway writes data/brain/learned.md, imports it into every coin as pad-lessons (versioned like the canon), fires a learned event, and quotes it at the top of every PAD.md. Pad admins (MIRRORPAD_BRAIN_ADMINS, wallets) can run it now with POST /brain/learn.

The main agent

A pad-level Mirror worker, role brain, workspace data/tokens/pad, read-only tools, system prompt: keeper of the shared brain, answers about VectorPad itself, never speaks for a coin. POST /brain/chat {message} asks it (guarded like a visitor turn, rate-limited per IP or wallet by MIRRORPAD_BRAIN_CHAT_RATE, 6 per 10 min); GET /brain shows its state. It is started on first use like any agent and reaped never (it is not builder or ops).

Knobs

Cost: one model turn per coin per day for reflection, one for the pad per day for lessons, about a tenth of a cent each on DeepSeek V3.2; canon imports are CLI calls, no model.

Verified

• Unit tests (tests/test_brain.py): rendering and placeholder safety, every canon file renders from settings and the version follows content, venue paragraphs per coin source, reflection parsing, themes, the brief, both digests, the routes (dev-only reflect, 503 not 500 without a worker, admin-only learn), versioned re-import of changed facts.
• On the mock model locally: a launched coin's community memory store held all nine canon facts and its official links within seconds of launch; PAD.md appeared with the coin's numbers and rank; the routes answered.
• Not yet run on a real model: a reflection and a lessons rewrite. gateway/scripts/agent_readiness.py is the place to add both once a model key is in the environment.