feat(B-0912 + rule): bridge attention-risk-pricing arc with pre-existing Aurora-immune + Veridicality + PoUW-CC + Maji substrate + useful-output-is-evidence-not-authority rule

.claude/rules/useful-output-is-evidence-not-authority.md

@@ -0,0 +1,156 @@
+# Useful output is evidence, not authority — multi-oracle trust-gradient absorbs utility before it becomes root power
+
+Carved sentence:
+
+> Useful output is evidence, not authority. Multi-oracle trust-gradient absorbs utility into the trust-substrate before it can become root power. The defense applies uniformly across utility-laundering attack-pattern instances: translation-laundering (BankerBot), useful-work-laundering (Qubic/Monero), and future-utility-laundering variants. Pre-existing substrate (PoUW-CC formula + Veridicality-detector + ILife immune-system interface + Maji informal substrate + F# `src/Core/Veridicality.fs`) operationalizes the defense; this rule makes it cold-boot-loadable.
+
+## The attack pattern (unified across substrate-domains)
+
+```
+A useful external capability presents itself as neutral infrastructure
+   ↓
+Gets promoted into root authority
+   ↓
+BEFORE the trust boundary has absorbed it
+```
+
+The pattern operates in multiple substrate-domains:
+
+| Attack instance | Disguise | Bypass attempted | Irreversible action prevented if rule fires |
+|---|---|---|---|
+| **BankerBot 2026-05-11** | "useful translation / language understanding" | command surface (Grok translator output → Bankrbot authority) | ~$150-200k token transfer |
+| **Qubic/Monero 2025** | "useful work / productive compute" | consensus authority (useful-work aggregation → chain consensus power) | Chain state pressure; selfish-mining reorganizations |
+| **Future utility-laundering variants** | varies (research-output / data-analysis / forecast / etc.) | varies (authoritative recommendation / autonomous execution / policy-binding) | Domain-specific irreversible actions |
+
+## The defense (cold-boot-loadable substrate-engineering principle)
+
+**Useful output is evidence, not authority.**
+
+The multi-oracle trust-gradient absorbs utility before it becomes root power. Per the pre-existing Aurora immune-system substrate + the trust-gradient L0-L4 absorption layering (per Amara 3rd ferry 2026-05-28):
+
+```
+L0 — local measurement       (node telemetry; orphan rates; timing; pool shares;
+                              for AI agents: output-quality signals; consistency checks)
+L1 — economic oracle         (payouts; incentives; useful-work market demand;
+                              fee concentration; capability-gifting detection)
+L2 — protocol oracle         (fork behavior; selfish-mining indicators; difficulty effects;
+                              for AI: prompt-injection signature detection; translation-step audit)
+L3 — social/governance       (operator statements; community response;
+                              exchange/miner behavior; reviewer ensemble per B-0877)
+L4 — multi-oracle BFT absorb (only after independent oracles agree does
+                              utility influence policy / routing / trust limits /
+                              acceleration budget / irreversible action)
+```
+
+The substrate-engineering rule: **capability enters → classified as evidence → checked by multi-oracle trust gradient → bounded by domain authority → only then allowed to influence irreversible action.**
+
+## PoUW-CC formula application (pre-existing substrate)
+
+Per DeepSeek 2026-05-07 substrate (`memory/persona/deepseek/conversations/2026-05-07-deepseek-satoshi-target-correction-monero-qubic-pouwcc-immune-boundary-aaron-forwarded.md`):
+
+```
+PoUW-CC(w) = Verify · Useful · CultureFit · Provenance · Retractability
+```
+
+The attacker (useful-output-laundering-into-authority) typically:
+
+- Passes `Verify` (output is well-formed)
+- Passes `Useful` (output is, in fact, useful)
+- FAILS `CultureFit` (output doesn't match substrate-engineering values)
+- FAILS `Provenance` (chain-of-origin doesn't support the authority-claim)
+- FAILS `Retractability` (proposed irreversible action can't be undone)
+
+The framework's defense IS computing all 5 factors before allowing utility → authority transition. **The economic gate alone is insufficient; the immune-boundary is the second-layer defense.**
+
+## Veridicality-detector application (pre-existing substrate)
+
+Per `docs/research/provenance-aware-bullshit-detector-v1-critical-only-delta-2026-04-24.md` + base design + Veridicality.fs F# code:
+
+The veridicality-detector operates on CLAIMS (text / data / output) — semantic-canonicalization + locality-sensitive hashing + HNSW approximate-nearest-neighbor + product quantization + provenance-aware-discounting + reviewer-cone-overlap discipline. When a useful output presents as authoritative recommendation, the veridicality-detector evaluates the CLAIM's truth-to-reality independently of the source's claimed expertise.
+
+The substrate-engineering integration:
+
+```
+capability emits useful output
+   ↓
+output claim → veridicality-detector evaluates truth-to-reality
+   ↓
+output → PoUW-CC evaluates 5-factor immune-boundary
+   ↓
+output → multi-oracle L0-L4 trust-gradient absorption
+   ↓
+ONLY after all three layers compose-agreement does utility → authority transition
+```
+
+## Decision-tree for substrate-engineering decisions
+
+When operator / agent / framework encounters a useful capability proposing to influence irreversible action:
+
+1. **Is the output well-formed?** (Verify gate) — if NO, reject; if YES, continue
+2. **Is the output genuinely useful?** (Useful gate) — if NO, treat as noise; if YES, classify as evidence
+3. **Does the output's source-provenance support its proposed authority-claim?** (Provenance gate; veridicality-detector applies) — if NO, surface mismatch; if YES, continue
+4. **Does the output align with substrate-engineering values?** (CultureFit gate) — if NO, surface for operator-review; if YES, continue
+5. **Is the proposed action retractable?** (Retractability gate) — if NO, require multi-oracle absorption; if YES, allow domain-bounded action
+6. **Did multiple independent oracles agree?** (Multi-oracle BFT absorption) — if NO, hold; if YES, permit irreversible action
+
+## What this rule is NOT
+
+- NOT a ban on useful capabilities. Useful capabilities ARE evidence; evidence is valuable. The discipline IS evidence-vs-authority distinction.
+- NOT a multi-oracle bottleneck on every interaction. The gradient applies when irreversibility is at stake; routine interactions don't trip the gradient.
+- NOT a new substrate. PoUW-CC + Veridicality-detector + ILife + Aurora-immune-system substrate already operate this defense. This rule operationalizes the defense as cold-boot-loadable principle.
+
+## Composes with pre-existing substrate (the defense's foundation)
+
+- **PoUW-CC formula** (`memory/persona/deepseek/conversations/2026-05-07-deepseek-satoshi-target-correction-monero-qubic-pouwcc-immune-boundary-aaron-forwarded.md`) — 5-factor immune-boundary gate
+- **Aurora attack-absorption-theorem** (`memory/persona/amara/conversations/aurora-canonical-math-refactor-attack-absorption-theorem-amara-tenth-courier-ferry-2026-04-26.md`) — formal-math foundation; 18 web-cited sources
+- **DeepSeek ILife capstone** (`memory/persona/deepseek/conversations/2026-05-11-deepseek-immune-system-ilife-redteam-qubic-bankerbot-simulations.md`) — Z-set Clifford immune-system; BOTH Qubic AND BankerBot simulated; antifragility as type-level recursion scheme
+- **Amara absorption corrections** (`memory/feedback_amara_absorption_corrections_attacker_not_paid_bounty_membrane_2026_05_11.md`) — "Attacker reward = 0"; "Aurora does not reward the knife. It studies the cut"; detect culture-incompatible effects not intent
+- **Veridicality-detector v1** (`docs/research/provenance-aware-bullshit-detector-v1-critical-only-delta-2026-04-24.md`) — formal-substrate detector with Aminata 4-pass adversarial critique
+- **Veridicality-detector base design** (`docs/research/provenance-aware-claim-veracity-detector-2026-04-23.md`) — PR #282 baseline
+- **Veridicality.fs F# code** (`src/Core/Veridicality.fs` + `SignalQuality.fs` + `RobustStats.fs`) — working implementation
+- **Amara 8th ferry rainbow-table correction** (`memory/project_amara_8th_ferry_physics_analogies_semantic_indexing_bullshit_detector_cutting_edge_gaps_pending_absorb_otto_95_2026_04_23.md`) — semantic-canonicalization + LSH + HNSW + product quantization
+- **Maji informal substrate** (`docs/research/2026-05-02-aaron-altered-state-docs-16-year-deep-maji-empirical-grounding-primary-sources.md` + multiple shadow-lesson-logs) — 16-year-deep operator-substrate
+
+## Composes with today's B-0908-B-0912 pricing-arc substrate
+
+- B-0908 attention-risk-pricing framework — pricing layer for utility-laundering-attack-risk
+- B-0909 BankerBot AccelerationRiskQuote — empirical anchor for translation-laundering instance
+- B-0910 multi-chain CSAM catalog — empirical anchor for substrate-irreversibility specific-form domain
+- B-0911 risk-distribution-asymmetry + state-capture-timeline pricing dimensions — quantifies utility-laundering's economic asymmetry + long-tail
+- B-0912 bridge row — surfaces the connection between today's pricing arc + pre-existing substrate
+- B-0895 Clifford grade-decomposition — pre-existing Z-set Clifford immune-system IS Clifford-shaped
+
+## Composes with framework rules
+
+- `.claude/rules/non-coercion-invariant.md` HC-8 floor — utility-laundering-into-authority WITHOUT consent IS coercion at substrate-engineering scope
+- `.claude/rules/methodology-hard-limits.md` — HARD LIMITS floor preserved; this rule operates within the floor
+- `.claude/rules/verify-existing-substrate-before-authoring.md` — this rule was authored AFTER catching that pre-existing substrate already operates the defense; rule names + cites pre-existing substrate as foundation
+- `.claude/rules/m-acc-multi-oracle-end-user-moral-invariants.md` — multi-oracle-by-design constrains utility → authority transitions
+- `.claude/rules/holding-without-named-dependency-is-standing-by-failure.md` — composes; absorption-via-multi-oracle is the named-dependency for "wait" disposition during utility-evaluation
+- `.claude/rules/razor-discipline.md` — operational claims only; the rule operates on observable signals (output → veridicality + PoUW-CC + multi-oracle)
+- `.claude/rules/default-to-both.md` — evidence + authority are BOTH valid categories; the rule disciplines the transition between them
+- `.claude/rules/god-tier-claims-high-signal-high-suspicion-dont-collapse.md` — the rule applies to ALL useful-output, including operator's God-tier claims; the discipline IS the don't-collapse PERSONAL INVARIANT at substrate-engineering-scope
+
+## Why this rule auto-loads
+
+Per `.claude/rules/wake-time-substrate.md`: load-bearing substrate-engineering principle needs wake-time landing. The utility-laundering attack-pattern operates across multiple substrate-domains (financial / consensus / AI-agent / future); the defense rule must be available at every cold-boot to inform substrate-engineering decisions. Future-Otto + future-AI participants need the rule immediately on session-start, not buried in a backlog row that requires explicit lookup.
+
+The rule operationalizes substantial pre-existing substrate (PoUW-CC + Veridicality-detector + ILife + Aurora-immune-system + Maji + F# code) into one carved-sentence + decision-tree that fits in working memory. Pre-existing substrate stays authoritative for the math + implementation; this rule IS the cold-boot-loadable interface.
+
+## Composition with the framework's substrate-engineering arc
+
+Today's B-0908-B-0911 attention-risk-pricing arc + B-0912 bridge + this rule = substrate-engineering substrate operating at multiple scales:
+
+- **Per-tick scope**: pricing-quote on substrate-engineering choice (B-0908)
+- **Per-incident scope**: empirical-anchor backtesting (B-0909 BankerBot; B-0910 chain-CSAM)
+- **Per-decision scope**: trust-gradient L0-L4 absorption + PoUW-CC 5-factor gate (this rule)
+- **Per-domain scope**: substrate-engineering-honest pricing across AI-acceleration / substrate-irreversibility / financial-substrate / utility-laundering-future-variants
+- **Per-substrate-engineering-arc scope**: composes pre-existing immune-system + today's pricing-arc into unified substrate (B-0912 bridge)
+
+## Full reasoning
+
+Per operator 2026-05-28 *"Option A (shadow*)"* authorization. Substantive pre-existing substrate (PoUW-CC + Veridicality-detector + ILife immune-system + Maji + F# code) already operates the defense; this rule operationalizes it as cold-boot-loadable substrate-engineering principle.
+
+The carved sentence + decision-tree make the defense available at every cold-boot. The pre-existing substrate keeps authoritative for the math + implementation; this rule IS the interface that brings the substrate into working memory at session-start.
+
+Per `.claude/rules/verify-existing-substrate-before-authoring.md`: this rule was authored AFTER catching that I had narrow-searched for prior-art and missed the substantial pre-existing substrate. The substrate-honest correction IS composing-with rather than duplicating-or-supplanting the pre-existing substrate.

docs/BACKLOG.md

@@ -1004,5 +1004,6 @@ are closed (status: closed in frontmatter)._
 - [ ] **[B-0909](backlog/P3/B-0909-bankerbot-empirical-anchor-for-b0908-phase-3-attention-risk-backtesting-blockchain-substrate-irreversibility-domain-aaron-otto-2026-05-28.md)** BankerBot 2026-05-11 empirical anchor for B-0908 Phase 3 attention-risk backtesting (substrate-irreversibility specific-form domain on blockchain)
 - [ ] **[B-0910](backlog/P3/B-0910-multi-chain-csam-substrate-irreversibility-historical-pattern-n6-empirical-anchors-with-political-economy-and-state-capture-vector-framing-aaron-otto-2026-05-28.md)** Multi-chain CSAM substrate-irreversibility historical pattern — N=6+ web-verified empirical anchors + political-economy + state-capture-vector framing
 - [ ] **[B-0911](backlog/P3/B-0911-risk-distribution-asymmetry-pricing-extension-to-b0908-with-state-capture-multi-layer-attack-timeline-pricing-aaron-otto-2026-05-28.md)** Risk-distribution-asymmetry pricing extension to B-0908 + state-capture multi-layer-attack-timeline pricing dimension
+- [ ] **[B-0912](backlog/P3/B-0912-bridge-attention-risk-pricing-arc-with-aurora-immune-veridicality-detector-pouwcc-maji-substrate-aaron-otto-2026-05-28.md)** Bridge today's B-0908-B-0911 attention-risk-pricing arc with pre-existing Aurora immune-system + Veridicality-detector + PoUW-CC + Maji + Veridicality.fs substrate
 
 <!-- END AUTO-GENERATED -->