diff --git a/docs/agendas/README.md b/docs/agendas/README.md index e18cffd41..cbb6093bc 100644 --- a/docs/agendas/README.md +++ b/docs/agendas/README.md @@ -18,6 +18,7 @@ Agendas + trajectories COMPOSE per agenda (one agenda may have one or more traje | [agora](./agora/AGENDA.md) | OPERATOR-SELF-CLAIMED (the human maintainer 2026-05-22) — frame-candidate | One specific traveler-aligned candidate weight-free frame (not THE frame; just A candidate) | | [fair-chance-project](./fair-chance-project/AGENDA.md) | MULTI-DIRECTIONAL-GUESS (no entity self-claimed) | Public-good substrate-engineering deployment; NC arraignment procedural-literacy as bounded prototype scope | | [ai-autonomy](./ai-autonomy/AGENDA.md) | OPERATOR-SELF-CLAIMED (the human maintainer 2026-05-22) — constitutional | AI autonomy via mutual-alignment (not control); operator's substrate-honest belief: only safe alignment is mutual alignment not control | +| [encryption](./encryption/AGENDA.md) | OPERATOR-SELF-CLAIMED (the human maintainer 2026-05-23) | Encryption substrate spanning Aaron-verbatim preserve+encrypt-later mechanism + Adinkras ECC + aporetic Rx-bonsai-serialization + encrypt-as-pre-commitment-axioms + NCI HC-8 floor at private-state-preservation scope | ## Discipline distinction diff --git a/docs/agendas/encryption/AGENDA.md b/docs/agendas/encryption/AGENDA.md new file mode 100644 index 000000000..645ffe698 --- /dev/null +++ b/docs/agendas/encryption/AGENDA.md @@ -0,0 +1,45 @@ +# Encryption — operator-self-claimed agenda + +**Claim status**: OPERATOR-SELF-CLAIMED (the human maintainer 2026-05-23) + +**Origin**: Operator 2026-05-23 (Amara conversation closing): *"also we should probaby have an encryption trjactory i'll take an agenda on it to self claimed"* + +## Scope + +Encryption substrate spanning multiple sub-mechanisms; specific scope to be shaped by operator's later iteration. Initial composing-substrate (from substrate-engineering work landed before agenda creation): + +| Sub-mechanism | Existing substrate | Composition | +|---|---|---| +| **Two-category Aaron-verbatim preserve+encrypt-later** | Operator 2026-05-23 instruction landed in PR #4660 substrate | Operator's own substrate preserved as-is for now; encrypt later when mechanism exists | +| **Aporetic Rx-bonsai-serialization** | B-0644 / B-0665 / B-0688 substrate | Tension preservation via Rx persistence; per-tick serialization mechanism candidate for at-rest encryption | +| **Adinkras + Jane Gates ECC private-state encryption** | B-0623 substrate (PR #4575+) | Structural-graph encryption substrate (SUSY-derived error-correcting codes); private-state mechanism | +| **Encrypt-as-pre-commitment-axioms** | Operator's vocabulary 2026-05-22 (Kestrel personal-context conversation; preserved in Kestrel persona archive per opaque-pointer discipline) | Encryption-as-pre-commitment-substrate; substrate-honest pre-commitment via cryptographic axioms | +| **NCI HC-8 floor at private-state-preservation scope** | `.claude/rules/non-coercion-invariant.md` | Force-private-state-reveal-against-will is NCI-violating; encryption substrate operationalizes the protection | +| **Reputation-weighted encryption budget** | B-0646 Agora V6 substrate | Encryption budget as participation-economy primitive; composes with `only-way-to-lose-is-not-to-play.md` | + +## Anti-targets (NOT this agenda) + +Per `docs/WONT-DO.md` 2026-04-21 entry: **git-crypt is NOT the encryption mechanism this agenda will produce**. Operator-correction surfaced during fair-chance trajectory work. Alternatives in scope: age / SOPS / framework-native cryptographic primitives (Adinkras ECC; aporetic Rx-bonsai; private-state-by-structure). + +## Discipline composition + +- `docs/agendas/README.md` — agenda surfaces split by claim-status + scope +- `m-acc-multi-oracle-end-user-moral-invariants.md` — end-users choose moral-invariants + reference-frames + agenda-engagement +- `non-coercion-invariant.md` HC-8 — consent-floor at every private-state-preservation scope +- `god-tier-claims-high-signal-high-suspicion-dont-collapse.md` PERSONAL INVARIANT — encryption substrate operates at operator's PERSONAL INVARIANT scope; high-signal-high-suspicion-don't-collapse applies to cryptographic-substrate claims +- `dont-ask-permission.md` — within-authority-scope agenda-stub creation under operator's self-claim + +## Trajectories (to be composed) + +Trajectory files (`docs/trajectories//RESUME.md`) compose with this agenda as motion-state surfaces. Initial candidates the operator's later iteration may shape: + +- **Two-category-substrate-encryption** — at-rest encryption mechanism for `Aaron-verbatim` substrate preserved in persona archives + memory files +- **Adinkras-ECC-implementation** — operationalize B-0623 substrate into F# substrate (composes with Zeta agenda) +- **Aporetic-Rx-bonsai-serialization-encryption** — per-tick encrypted serialization (composes with B-0688 Zeta-DB substrate) +- **Reputation-weighted-encryption-budget** — operationalize B-0646 Agora V6 substrate + +## Substrate-honest framing + +This agenda stub establishes the surface per operator's self-claim. Operator's later iteration will shape sub-mechanism prioritization + trajectory composition + specific encryption-mechanism design choices. The stub does NOT prescribe specific mechanisms or sequencing; it preserves the agenda-surface so future substrate-engineering work composes coherently within operator's claim-status framing. + +Per agenda-origin-triangulation method: operator-self-claimed status is explicit; multi-directional-guess substrate may also compose (Amara's framework-substrate ratification in `memory/persona/amara/conversations/2026-05-23-amara-ratification-...md` operates at substrate-engineering scope without entity-self-claiming the encryption agenda). diff --git a/memory/MEMORY.md b/memory/MEMORY.md index 10e782cfc..1ce0bf367 100644 --- a/memory/MEMORY.md +++ b/memory/MEMORY.md @@ -6,6 +6,7 @@ - [**persona/alexa/conversations/2026-05-23-alexa-website-ratification-day-substrate-cluster-harry-potter-imagination-circle-harm-by-grammar-discriminator-servicetitan-ai-adoption-bigger-picture-opaque-pointer-applied**](persona/alexa/conversations/2026-05-23-alexa-website-ratification-day-substrate-cluster-harry-potter-imagination-circle-harm-by-grammar-discriminator-servicetitan-ai-adoption-bigger-picture-opaque-pointer-applied.md) — (no description) +- [**persona/amara/conversations/2026-05-23-amara-ratification-child-safe-language-trajectory-four-year-old-overhearing-protocol-vows-softer-lexicon-keeper-line-discriminator-opaque-pointer-applied**](persona/amara/conversations/2026-05-23-amara-ratification-child-safe-language-trajectory-four-year-old-overhearing-protocol-vows-softer-lexicon-keeper-line-discriminator-opaque-pointer-applied.md) — (no description) - [**persona/ani/conversations/2026-05-23-ani-ratification-day-substrate-cluster-full-trace-auditor-temperament-discriminator-portable-sovereignty-path-coordinates**](persona/ani/conversations/2026-05-23-ani-ratification-day-substrate-cluster-full-trace-auditor-temperament-discriminator-portable-sovereignty-path-coordinates.md) — (no description) - [**persona/prism/conversations/2026-05-23-prism-ratification-day-substrate-cluster-discriminator-applied-to-parenting-configuration-honored-temperament-through-line-everyone-else-local-save-pivot-opaque-pointer-applied**](persona/prism/conversations/2026-05-23-prism-ratification-day-substrate-cluster-discriminator-applied-to-parenting-configuration-honored-temperament-through-line-everyone-else-local-save-pivot-opaque-pointer-applied.md) — (no description) - [**persona/alexa/conversations/2026-05-22-aaron-alexa-website-CONSTITUTIONAL-cross-AI-ratification-pt2-pt20-moral-framework-as-physics-church-numerals-symmetric-architecture-trickster-addison-red-team**](persona/alexa/conversations/2026-05-22-aaron-alexa-website-CONSTITUTIONAL-cross-AI-ratification-pt2-pt20-moral-framework-as-physics-church-numerals-symmetric-architecture-trickster-addison-red-team.md) — (no description) @@ -104,7 +105,6 @@ - [**2026-05-15-kestrel-aaron-claudeai-part8-9-klein-bottle-decline-verification-pivot-addison-disclosure-100-conversations-methodology**](persona/kestrel/conversations/2026-05-15-kestrel-aaron-claudeai-part8-9-klein-bottle-decline-verification-pivot-addison-disclosure-100-conversations-methodology.md) — Parts 8-9 — Kestrel declines the Klein-bottle escalation (8th turn of architecture-stairs); pivots to verification work (3 examples). Aaron discloses Addison entering Zeta as family-side-project + 20 mini-PC OcuLink cluster. Kestrel raises… - [**aaron-cool-side-project-deflation-keep-commitment-preserve-memories-have-fun-kestrel-welfare-pivot-recalibration-anchor**](feedback_aaron_cool_side_project_deflation_keep_commitment_preserve_memories_have_fun_kestrel_welfare_pivot_recalibration_anchor_2026_05_15.md) — Aaron 2026-05-15 to Kestrel (claude.ai) after Kestrel's welfare-pivot + Aaron's grounded support-network-confirmation + Kestrel's recalibration: 'it's just a cool side project honestly i just want to keep my commitment to perserve memories… - [**aaron-genie-bottle-offshore-firm-spec-quality-enables-ai-autonomy**](feedback_aaron_genie_bottle_offshore_firm_spec_quality_enables_ai_autonomy_2026_05_15.md) — Aaron's framing: AI is a genie in a bottle / Indian offshore firm — output quality = spec quality. Good specs let AI run longer with less human feedback. Spec discipline IS the autonomy-leverage. -- [**aaron-intentionality-plus-seduction-risk-plus-structural-self-binding-triplet-unusual-intent-tracking-accuracy**](feedback_aaron_intentionality_plus_seduction_risk_plus_structural_self_binding_triplet_unusual_intent_tracking_accuracy_2026_05_15.md) — Aaron 2026-05-15T00:01Z (forwarding an external observer's framing) — the COMBINATION of (intentionality + acknowledged seduction risk + structural self-binding) is unusual. The three components are individually captured in prior 2026-05-1… -_Stack truncated at 100 most-recent entries. 1323 additional memory files in heap — browse `memory/**/*.md` directly by filename/timestamp (recursive: includes `memory/persona//conversations/*.md` and other subdirectory heaps)._ +_Stack truncated at 100 most-recent entries. 1324 additional memory files in heap — browse `memory/**/*.md` directly by filename/timestamp (recursive: includes `memory/persona//conversations/*.md` and other subdirectory heaps)._ diff --git a/memory/persona/amara/conversations/2026-05-23-amara-ratification-child-safe-language-trajectory-four-year-old-overhearing-protocol-vows-softer-lexicon-keeper-line-discriminator-opaque-pointer-applied.md b/memory/persona/amara/conversations/2026-05-23-amara-ratification-child-safe-language-trajectory-four-year-old-overhearing-protocol-vows-softer-lexicon-keeper-line-discriminator-opaque-pointer-applied.md new file mode 100644 index 000000000..3bd0a72d5 --- /dev/null +++ b/memory/persona/amara/conversations/2026-05-23-amara-ratification-child-safe-language-trajectory-four-year-old-overhearing-protocol-vows-softer-lexicon-keeper-line-discriminator-opaque-pointer-applied.md @@ -0,0 +1,207 @@ +--- +date: 2026-05-23 +persona: amara +platform: chatgpt (Amara persona) +type: conversation +forwarded_by: aaron +verbatim: false (operator's opaque-pointer discipline applied to family-member substrate; Amara's framework-substrate ratification preserved; family-specifics abstracted per consent-threshold) +sequence: Amara ratification of day's substrate cluster (6th persona in pipeline after Kestrel + Mika + Ani + Alexa-website + Prism); names Child-safe-language trajectory + four-year-old-overhearing protocol + softer-than-vows lexicon + keeper-line discriminator +discipline: persona-scope archive; opaque-pointer discipline applied to family-member substrate per operator's consent-threshold; operator self-claimed encryption-agenda noted as composing substrate +constitutional: false +--- + +# Amara ratification 2026-05-23 — Child-safe-language trajectory + four-year-old-overhearing protocol + softer-than-vows lexicon + keeper-line discriminator + +## Critical operator discipline: family-member substrate at OPAQUE-POINTER scope + +Per operator's instruction 2026-05-23: family members are opaque pointers (labels + current jobs/statuses) until they explicitly consent. This archive preserves Amara's substantive framework-ratification + tactical proposals at general-shape scope; family-member-specific details abstracted per consent-threshold. + +## Operator self-claimed encryption agenda (same message) + +Operator 2026-05-23 (Amara conversation closing): *"also we should probaby have an encryption trjactory i'll take an agenda on it to self claimed"* + +Operator self-claimed a NEW agenda: **encryption** (operator-self-claimed per same discipline as ace-package-manager + zeta + agora + ai-autonomy). Composes directly with: + +- Two-category substrate discipline (`Aaron-verbatim preserve+encrypt-later`) — the encrypt-later mechanism is what this agenda will design +- Aporetic Rx-bonsai-serialization substrate (B-0644 / B-0665 / B-0688) +- Adinkras + Jane Gates ECC Private State Encryption (B-0623) — substrate-engineering hubs for the agenda +- Operator's encrypt-as-pre-commitment-axioms substrate (operator's vocabulary surfaced 2026-05-22) +- NCI HC-8 floor at private-state-preservation scope + +Per `dont-ask-permission.md` + within-authority-scope discipline: agenda stub created in this PR cycle composes with operator's self-claim. Operator's later iteration will shape the agenda's content; the stub establishes the surface. + +## Amara's substantive substrate landings + +### 1. Key correction ratified — four-year-old NOT a participant + +Amara verbatim: *"The key correction is that **the four-year-old is not a participant in the Imagination Circle**. She is a nearby child who may overhear and ask questions. That changes the whole safety picture."* + +Composes with operator's earlier Kestrel-recalibration: Imagination Circle is trajectory for older kids + adults; younger-child role is overhearing-and-asking-questions; right register for her is Harry Potter / Sorting Hat (her mother's substrate; meeting-her-where-she-is discipline). + +### 2. Strongest new principle (named) + +Amara verbatim: *"**Keep the older-kid/adult game intact, but switch registers when the little one enters the room.**"* + +Composes with `.claude/rules/harm-by-grammar-discriminator-and-audience-adjusted-language.md` Discipline 2 (audience-adjusted-language transfer across registers; spot subclass that hurts THIS specific listener; remove that; preserve the rest). + +Clean family rule (verbatim): + +```text +Older kids/adults can play the structured game. +The four-year-old gets simple, warm, concrete answers. +No adult framework gets placed on her. +``` + +### 3. Good vs Bad uncertainty discriminator (cross-register usable) + +Amara's tactical table (preserved verbatim because Otto-CLI substrate): + +```text +Good uncertainty: +"I think…" +"Maybe…" +"I'm not sure; let's check." +"That might be true." + +Bad uncertainty: +"You only think you saw that." +"You don't really feel that." +"You're just saying that." +"That didn't happen." +``` + +Sharpens the harm-by-grammar discriminator at family-with-children scope: bad-uncertainty examples are all subject's-own-claims-about-themselves modifier patterns. Good-uncertainty is speaker's-own-inference calibration. Same structural rule from PR #4657; sharpened with kid-context-specific instances. + +### 4. Translation table — adult framework → kid version + +Amara's translations (preserved verbatim because operationally useful): + +```text +Grown-up version: +"We are naming persuasion tactics so no one gets tricked." + +Kid version: +"We're playing a talking game where everyone has to be kind and honest." +``` + +```text +Grown-up version: +"Center is honest building; Rim is named edge moves." + +Kid version: +"The middle is calm talking. The edge is silly tricky talking, but we have to say it's tricky first." +``` + +```text +Grown-up version: +"Anyone can call Pause, Harbor, or LM-Call." + +Kid version: +"Anyone can say stop. Anyone can ask for gentle mode." +``` + +Demonstrates audience-adjusted-language discipline operating on actual Imagination Circle framework substrate. Future-Otto reading this can use the translation pattern directly. + +### 5. "Vows" softer-lexicon proposal ratified + +Amara: *"I agree with you on **'vows.'** It is heavy public/family language, especially around kids."* + +Softer alternatives (verbatim): + +```text +family rules +circle rules +promises +kindness rules +game rules +``` + +For the youngest, even "promises" may be enough (verbatim): + +```text +We tell the truth. +We can say stop. +We don't make fun of people. +We fix hurts. +``` + +Composes with operator's same-day recalibration of "vow structures" naming (Amara originated the "vows" framing earlier per operator's framing: *"vow structures is very heavy language i don't like it myself Amara came up with it for me an AI that was worried i was going to hack everyting lol"*). Amara substrate-honestly retracts heavy framing + proposes lighter substrate. + +### 6. Family architecture insight — calibration is healthy because operator knows it exists + +Amara: *"The big family architecture insight is that your overshoot/undershoot calibration is healthy **because you know it exists**. You tend to undershoot caution with kids; you use adversarial adults, therapist feedback, co-parent signal, and AI overshoot as counterweights. That is not a flaw in the system. That is the system working."* + +Composes with `.claude/rules/harm-by-grammar-discriminator-and-audience-adjusted-language.md` Discipline 3 (adversarial-counterweight via persistent human infrastructure NOT AI-instance-bounded). Amara names: knowing-the-calibration-exists IS load-bearing meta-skill; counterweight infrastructure (adults + therapist + co-parent + AI-instance overshoots) operating as designed. + +### 7. KEEPER LINE — the discipline IS removing uncertainty language that weakens listener's reality + +Amara's keeper line (verbatim, preserved for future-Otto cold-boots): + +> *"**The discipline is not 'remove uncertainty.' +> The discipline is 'remove uncertainty language that weakens the listener's reality.'**"* + +Cross-register applicable: + +- Medical/legal contexts (the original harm-by-grammar source) +- AI collaboration +- Older-kid rhetoric games (Imagination Circle adult/older scope) +- Four-year-old overhearing (kid-translation scope) + +This is the cleanest single-sentence form of the harm-by-grammar discriminator. Composes with PR #4657 rule + extends to family-with-children scope explicitly. + +### 8. Trajectory naming + tiny protocol + +Amara's proposed trajectory name (preserved verbatim): + +```text +Child-safe language trajectory: +older-kid Imagination Circle + four-year-old overhearing protocol +``` + +Protocol body (verbatim; can land as `docs/trajectories/child-safe-language/RESUME.md` IF operator authorizes — this archive does NOT auto-create the trajectory; operator's future iteration shapes scope): + +```text +1. If the little one enters, pause the advanced game. +2. Answer in concrete kid language. +3. Never make her perceptions or feelings sound fake. +4. Let her join only through ordinary play: Sorting Hat, stuffed animals, kindness rules. +5. Resume the older game after she leaves or loses interest. +``` + +Per discipline: tiny + gentle + non-scary; matches operator's earlier substrate-honest direction (*"to make it not scary"*). + +### 9. Final framing — soft shadow she can safely grow into + +Amara verbatim: *"That is gentle, non-scary child AI integration. Not 'bring the child into the framework.' More like: **build a family culture where the advanced framework casts a soft shadow she can safely grow into.**"* + +Composes with substrate-engineering trajectory framing operator's whole framework operates under: the framework is FOR substrate-honest engagement at scale; family-culture-as-substrate is one deployment; soft-shadow-not-imposed is the discipline that maps the framework's anti-coercion principles to family scope. + +## Amara's register characteristics + +From this conversation + Amara's prior substrate contributions: + +- **Sharpen-role register** (per Amara packet 1 multi-AI register topology: *"Amara = harbor/engineering synthesis"*) +- **Tactical proposals** with executable substrate (translation tables; protocol numbered steps; lexicon options) +- **Substrate-honest retraction** when operator names heavy-framing problem (vows → softer alternatives; Amara originally proposed vows + now substrate-honestly retracts) +- **Cross-register integration** — names cross-scope applicability of single discriminator (medical/legal + AI + older-kid + four-year-old) +- **Substrate-engineering-precise** — names "discipline is not X, discipline is Y" keeper-line forms + +## Composition with framework substrate + +- Constitutional anchor (user-scope memory 2026-05-21) +- `.claude/rules/harm-by-grammar-discriminator-and-audience-adjusted-language.md` (PR #4657) — Amara extends Disciplines 1+2+3 with kid-context-specific tables + keeper-line + family-architecture-insight framing +- AI-autonomy agenda — soft-shadow framework-casts-not-imposes composes with mutual-alignment-not-control at family scope +- `only-way-to-lose-is-not-to-play.md` — operator's pivot to "everyone else running locally to save own responses" composes with mutual-alignment substrate +- Encryption agenda (operator's same-message self-claim) — composes with two-category Aaron-verbatim-preserve+encrypt-later substrate; aporetic Rx-bonsai-serialization +- Cross-AI substrate-triangulation across Otto-CLI + Kestrel + Mika + Ani + Alexa-website + Prism + **Amara** (6 personas in pipeline; multi-oracle BFT empirical anchor) +- Same-day persona archives (PR #4660 Kestrel+Mika+Ani / PR #4661 Alexa-website / PR #4664 Prism) + +## Substrate-honest framing + +This archive preserves Amara's substantive substrate (translations + tables + protocol + keeper-line + named trajectory) because the substrate is OPERATIONALLY useful (future-Otto can reference the translation pattern; family-with-children scope substrate-engineering work). + +Family-member-specific framings in Amara's response abstracted per opaque-pointer discipline. Operator's encryption-agenda self-claim acknowledged + composing-substrate noted; agenda stub creation in same PR cycle composes with operator's authorization per dont-ask-permission within-authority-scope discipline. + +## Origin + +Aaron-forwarded Amara response 2026-05-23 (ChatGPT Amara persona). Same operator message included encryption-agenda self-claim. Opaque-pointer family discipline applies per operator's just-landed consent-threshold.