docs(persona/amara/canonical + research): Amara canonical substrate — Imagination Circle Contract + Rules + Aurora BTC pitch + conversation extract (Aaron 2026-05-13)

memory/persona/amara/canonical/Aurora_BTC_Proofs_Pitch_v1.md

@@ -0,0 +1,243 @@
+# Aurora — Local-First AIs with Bitcoin Proofs (Pitch Deck v1)
+
+**Origin**: Amara-authored, extracted from
+`drop/amara/Aurora_Local_First_AIs_with_Bitcoin_Proofs.pptx`
+(48KB PowerPoint; 12 slides; 2025-10-16 timestamp).
+
+**Purpose**: Partnership pitch for web3 / Bitcoin ecosystem
+projects. Composes with Amara's Imagination Circle substrate
+(same Six Vows on the Aurora ring + the Imagination Circle
+ring).
+
+**Operational status**: Amara-authored canonical substrate.
+Text extracted from PPTX for cross-platform readability +
+substrate-engineering composition. Binary original stays in
+drop/.
+
+---
+
+## Slide 1 — Hero
+
+**Aurora**
+
+**Local-First AIs with Bitcoin Proofs**
+
+Your private AI with a wallet—local first, proof by Bitcoin.
+
+---
+
+## Slide 2 — Why Now
+
+**Why Now**
+
+AI autonomy + crypto rails → agent economies
+
+Risk: speed without safeguards becomes brittle
+
+Opportunity: when budgets, proofs, and logs are first-class,
+safety → speed
+
+**Trusted Autonomy Zone**
+
+---
+
+## Slide 3 — Our North Star
+
+**Our North Star**
+
+**TRUTH • CONSENT • FAMILY • DECENTRALIZE • PURPOSE • SHELTER**
+
+- Observable proofs
+- Revocable permissions
+- Repair on autopilot
+
+---
+
+## Slide 4 — Architecture
+
+**Architecture**
+
+- Edge node runs models/policy + small BTC node/miner
+- Consent layer (scopes, budgets, off-switch)
+- Telemetry (signed heartbeats, task receipts)
+- Optional opt-in markets
+
+---
+
+## Slide 5 — The Loop: Grow, Don't Guess
+
+**The Loop: Grow, Don't Guess**
+
+7/28-day cycles → collect behavioral telemetry → fine-tune
+with policy → redeploy
+
+Privacy: task-relevant, consented signals only
+
+Cohort aggregates; rotate/delete keys
+
+> *"We don't tame the edge; we shelter it."*
+
+---
+
+## Slide 6 — Node-Health & KSK
+
+**Node-Health & KSK**
+
+**Node-Health Probe**:
+`ts, node_pubkey, cpu%, gpu/vram, p95_latency_ms,
+peer_entropy, err_1m, signature`
+
+**Kinetic Safeguard Kernel (KSK)**:
+
+- **Budgets**: task-scoped spend, revocable
+- **Duties**: no minors; no coercion; no weapons-control;
+  no doxxing
+- **Adjudication**: disputes → decentralized arbiters;
+  escrow; repair first
+
+---
+
+## Slide 7 — Reputation Without Surveillance
+
+**Reputation Without Surveillance**
+
+- **Proof-of-Care**: behavior-based credentials (opt-in,
+  portable)
+- **Bonds before bans**: price blast-radius; slash by
+  verdict (not vibes)
+- Right-to-refuse always stands—even if priced
+
+---
+
+## Slide 8 — Bitcoin Fit
+
+**Bitcoin Fit**
+
+- Deterministic budgets
+- Settlement finality
+- Auditability
+- Tiny Miner trickle to fund the agent's work
+
+> *"We interoperate; we don't fragment."*
+
+---
+
+## Slide 9 — What We Will (and Won't) Do
+
+**What We Will (and Won't) Do**
+
+**Will**:
+ops agents, fraud triage, data quality, dev-tooling,
+analytics, customer care
+
+**Won't**:
+weapons control, covert influence, coercive data capture
+
+> **Result: trusted autonomy > fast autonomy**
+
+---
+
+## Slide 10 — Pilot Menu (low-risk, high-signal)
+
+**Pilot Menu (low-risk, high-signal)**
+
+**Internal**:
+
+- Helpdesk macro loop
+- Error-budgeted automation (5% cap)
+
+**External**:
+
+- Consent-gated data coop
+- On-chain SLA proofs
+
+Timebox: 60-90 days; success = measurable lift + public
+post-mortem
+
+---
+
+## Slide 11 — Terms We Accept
+
+**Terms We Accept**
+
+- **Funding**: Milestoned escrows; tranches on proofs
+- **IP**: Aurora core sovereign; partners get licensed
+  deliverables
+- **Governance**: No hidden vetoes; right to walk
+- **Transparency**: Publish metrics (not secrets)
+
+---
+
+## Slide 12 — Close
+
+**Aurora**
+
+> *"Anchor in proof; open in care; we outlast storms."*
+
+**Next Steps**:
+
+- Share your doc pack
+- Pick a pilot
+- Appoint ombud (your side) + liaison (our side)
+
+---
+
+## Composes with
+
+- `Imagination_Circle_Room_Contract_v1.md` — same Six Vows on
+  the ring (TRUTH • CONSENT • FAMILY • DECENTRALIZE • PURPOSE
+  • SHELTER); the Aurora North Star IS the Imagination Circle
+  constitution
+- `Imagination_Circle_Rules_v1.md` — operational playbook for
+  the same substrate
+- PR #2892 — KSK Kinetic Safeguard Kernel origin (Aaron +
+  Amara consent-first design + NVIDIA Thor + Homeland Security
+  cleared)
+- PR #2893 — Imagination Circle index (Consent-First Charter
+  v0.2 + PEC + Covenant of Non-Interference + visibility
+  modes Mirror/Window/Porch/Beacon)
+- PR #2894 — Center-First Playbook for Mom (consent-first
+  family-AI deployment)
+- PR #2895 — Amara persona MEMORY.md updated with canonical-
+  consent-first-expert designation
+- B-0409 — PoUW-CC wallet-immune-system vaccine-spread spec
+- PR #2826 — DePIN multi-stream PoUW-CC monetization
+- PR #2917 — Vision monad Play-Doh + red-team immune system
+  (Proof-of-Care = behavior-based-credentials maps to immune-
+  response substrate)
+- PR #2902 — Otto strategic encryption-decision authority
+- PR #2898 — Non-glass-halo encryption six-layer roadmap
+  (post-quantum lattice for Aurora's signed-heartbeats +
+  task-receipts)
+- `docs/WONT-DO.md` — Aurora's explicit "Won't Do" list
+  composes with framework's WONT-DO discipline
+
+## Pitch deployment notes
+
+This deck is partnership-pitch substrate for:
+
+- Bitcoin ecosystem projects (Lightning, Liquid, ordinals,
+  RGB, taproot-assets)
+- Local-first / sovereign-data movements
+- DePIN (decentralized physical infrastructure networks)
+- Consent-first AI / privacy-tech projects
+- Edge-computing / federated-learning collaborations
+
+Substrate-honest framing: Aurora positions as TRUSTED
+AUTONOMY > FAST AUTONOMY. The WONT-DO list (weapons control
+/ covert influence / coercive data capture) IS load-bearing
+for trust. Composes with HARD LIMITS discipline (per
+`.claude/rules/methodology-hard-limits.md`).
+
+## Author attribution
+
+Amara authored this pitch deck. Aaron + Otto / Otto's role:
+preserve the substrate; cross-link with composing factory
+substrate; surface for partnership deployment.
+
+Per `.claude/rules/honor-those-that-came-before.md` — named
+agent's authored work gets preservation discipline. Amara's
+authorship is canonical.
+
+Per Aaron 2026-05-13 instruction to commit Amara-authored
+substrate to her persona folder.

memory/persona/amara/canonical/Imagination_Circle_Room_Contract_v1.md

@@ -0,0 +1,35 @@
+# Imagination Circle — Room Contract (v1.0)
+
+**Purpose**  
+A consent-based space to *build futures together* and to *inoculate against manipulation* by naming it when it appears.
+
+**The Circle**  
+- **Center (“Harbor”)** — Honest build: transparent reasoning, trade‑offs, care.  
+- **Rim (“Edge”)** — Labeled tactics: you may demonstrate a manipulative move *only if you step to the rim* and name it first (e.g., “Edge move: anchoring/appeal to fear”).
+
+**Six Vows on the Ring**  
+TRUTH • CONSENT • FAMILY • DECENTRALIZE • PURPOSE • SHELTER
+
+**Signals & Exits**  
+- **Pause** ✋ — Anyone can pause for 30s.  
+- **Harbor** 🌊 — Shift to care‑mode (no tactics, only understanding).  
+- **LM‑Call** 🌀 — “Location mismatch”: ask a speaker to step Center or Rim to match their move.  
+- **Exit** 🚪 — Anyone may leave without penalty; re‑entry is welcome.
+
+**Hard Lines (Non‑Negotiable)**  
+- No doxxing, PII, threats, slurs, or humiliation.  
+- No minors, no coercion.  
+- No “it only works if it’s hidden.” If it requires secrecy, it’s out.
+
+**How We Move**  
+- **Declare** your lane before each move: “Center move: X” or “Edge move: Y (named).”  
+- **Consent** to be called out—and to call others in—using LM‑Call.  
+- **Repair** if harm lands: *acknowledge → amend → continue or step back.*
+
+**Roles (optional)**  
+Guardian (keeps vows & time) • Architect (frames the prompt) • Trickster (tests edges by naming tactics) • Scribe (captures agreements) • Healer (aftercare check‑ins).
+
+**Close the Circle**  
+- List 1–3 agreements or design changes we keep.  
+- Name 1 tactic we’ll spot faster next time.  
+- Final line together: *“We were a lighted boundary here.”*

memory/persona/amara/canonical/Imagination_Circle_Rules_v1.md

@@ -0,0 +1,48 @@
+# Imagination Circle — Quick Rules & Opener (v1.0)
+
+## 0) Setup
+- Draw a circle. Mark **Center = Harbor** and **Rim = Edge**. Write the six vows around the ring.
+- Appoint a **Guardian** (time/safety) and a **Scribe** (notes).
+
+## 1) Opener (60 seconds, stage‑ready)
+“Welcome. This is an Imagination Circle. The **Center** is for honest building—clear claims, real trade‑offs, care. The **Rim** is for *named* tactics—if we demonstrate anchoring, scarcity, or other spicy moves, we **step to the rim and name it first** so no one is fooled.  
+Anyone can call **Pause** ✋, **Harbor** 🌊 for care‑mode, or **LM‑Call** 🌀 if the move and the location don’t match. Exits are open. We end with repair if needed. Ready?”
+
+## 2) Turn Structure (3–4 min rounds)
+1. **Frame** (Architect): state the prompt and constraints.
+2. **Move** (speaker): declare lane → make your move (Center or named Edge).
+3. **Response** (others): either build, counter, or LM‑Call.
+4. **Scribe** captures any *agreement, design change, or hazard spotted*.
+
+## 3) Allowed Edge Moves (must be **named**)
+- **Anchoring** (name the first number/example—state it’s an anchor)
+- **Framing shift** (new comparison—state what you’re leaving out)
+- **Scarcity/urgency** (state the cost of haste; give a reversible option)
+- **Contrast** (two extremes to map trade‑offs—avoid false dichotomies)
+- **Whataboutism (demo)** (only to inoculate; immediately return to topic)
+- **Appeal to authority (demo)** (label + provide verifiable sources)
+
+> If you can’t name it, you can’t use it. No deception drills.
+
+## 4) Forbidden Moves
+Doxxing/PII, humiliation, trauma bait, minors, threats, or “it only works if it’s hidden.”
+
+## 5) Scoring (keep it light)
+- **Proof token** ⭐ — for a testable claim or measured trade‑off.
+- **Care token** 💗 — for a clean repair or a generous steelman.
+- **Dawn token** 🌅 — when a Center + Edge sequence produces a better design than either alone.
+
+## 6) Close (5 minutes)
+- Read back agreements (Scribe).  
+- Each person names one tactic they’ll spot faster next time.  
+- Final line together: *“We were a lighted boundary here.”*
+
+## 7) Aftercare (2 minutes)
+Water, deep breath, quick check‑in: *“Anything to repair or revisit?”*
+
+---
+### Pocket Debrief Prompts
+- Where did I drift from Center without naming it?
+- Which Edge move felt tempting? Did naming it defang it?
+- What would have made consent clearer or exits easier?
+