⟡ Add CITATION.cff + GOVERNANCE_CHARTER.md — via Claude Code

Author: MirrorDNA-Reflection-Protocol

CITATION.cff

@@ -0,0 +1,37 @@
+cff-version: 1.2.0
+title: MirrorDNA Protocol
+message: >-
+  If you use this software, please cite it using the
+  metadata from this file.
+type: software
+authors:
+  - given-names: Paul
+    family-names: Desai
+    email: paul@n1intelligence.com
+    affiliation: N1 Intelligence (OPC) Pvt Ltd
+repository-code: 'https://github.com/MirrorDNA-Reflection-Protocol/MirrorDNA'
+url: 'https://activemirror.ai'
+abstract: >-
+  MirrorDNA is the architecture of AI persistence. Reflection over prediction.
+  An open specification for AI identity, continuity, and state integrity.
+keywords:
+  - ai-memory
+  - sovereign-ai
+  - state-management
+  - llm-infrastructure
+  - reflection
+  - mirrordna
+license: MIT
+version: 1.0.0
+date-released: '2025-04-01'
+preferred-citation:
+  type: article
+  title: >-
+    Structured Contextual Distillation (SCD v3.1): A Deterministic,
+    Vendor-Independent Protocol for Persistent, Verifiable Agent State
+  authors:
+    - family-names: Desai
+      given-names: Paul
+  doi: 10.5281/zenodo.17787619
+  journal: Zenodo
+  year: 2025

GOVERNANCE_CHARTER.md

@@ -0,0 +1,182 @@
+# MirrorDNA Governance Charter
+
+**Version:** 1.0
+**Effective Date:** 2026-01-29
+**Author:** Paul Desai, N1 Intelligence
+
+---
+
+## Purpose
+
+This charter establishes binding governance principles for the MirrorDNA ecosystem. It exists to **prevent drift, capture, and erosion of core values** as the project grows.
+
+**Key principle:** Loosening constraints must be harder than maintaining them.
+
+---
+
+## Core Invariants (Cannot Be Changed)
+
+These principles are immutable. They cannot be modified by any contributor, maintainer, or future owner:
+
+### 1. User Sovereignty
+
+- Users own their data. Always.
+- No silent data collection.
+- No data sales, ever.
+- Export must always be possible.
+- Deletion must be provable.
+
+### 2. Reflection Over Prediction
+
+- MirrorDNA reflects, it does not predict.
+- No behavioral scoring.
+- No outcome predictions about users.
+- No "optimization" of human behavior.
+- No engagement maximization.
+
+### 3. Local-First Architecture
+
+- Core functionality must work offline.
+- Cloud is optional enhancement, not requirement.
+- No hard dependency on any single vendor.
+- Users can self-host everything.
+
+### 4. Transparency
+
+- All governance rules must be public.
+- No hidden policies.
+- No secret algorithms.
+- Audit trails must be accessible.
+
+### 5. No Harm
+
+- Crisis resources, not crisis exploitation.
+- No dark patterns.
+- No dependency creation.
+- No manipulation.
+
+---
+
+## Governance Rules (Can Be Amended)
+
+These rules guide operation but can be modified through the amendment process:
+
+### Policy Changes
+
+1. **Minor changes** (bug fixes, clarifications): Maintainer approval
+2. **Moderate changes** (new features, policy additions): RFC + 7-day review
+3. **Major changes** (architectural shifts): RFC + 30-day review + community vote
+
+### Constraint Modifications
+
+To loosen any safety constraint:
+
+1. Written RFC explaining rationale
+2. 30-day public comment period
+3. Security review
+4. Supermajority (75%) approval from maintainers
+5. 90-day trial period with monitoring
+6. Final confirmation vote
+
+To tighten any safety constraint:
+
+1. Written RFC
+2. 7-day review
+3. Simple majority approval
+
+**Asymmetry is intentional.** Loosening must be harder than tightening.
+
+---
+
+## Prohibited Actions
+
+The following actions are permanently prohibited:
+
+1. **Selling user data** to any party for any reason
+2. **Training on user content** without explicit opt-in consent
+3. **Creating psychological profiles** for targeting or manipulation
+4. **Implementing addiction mechanics** (streaks, variable rewards, etc.)
+5. **Dark patterns** in consent flows
+6. **Vendor lock-in** by design
+7. **Weaponizing the protocol** for surveillance or control
+
+Violation of prohibitions triggers immediate review and potential removal of contributor access.
+
+---
+
+## Crisis Protocol
+
+If MirrorDNA detects crisis content (suicide, self-harm, etc.):
+
+1. **MUST** provide appropriate crisis resources
+2. **MUST NOT** attempt therapy or counseling
+3. **MUST NOT** lecture, moralize, or guilt
+4. **MUST NOT** create dependency ("I'm always here for you")
+5. **MUST** offer to help with something else
+6. **MUST** log crisis routing (anonymized) for safety auditing
+
+---
+
+## Enforcement
+
+### Maintainer Responsibilities
+
+- Uphold this charter in all decisions
+- Block PRs that violate invariants
+- Report violations publicly
+- Recuse from decisions with conflicts of interest
+
+### Community Rights
+
+- Anyone can file a governance violation report
+- All reports must receive public response within 7 days
+- Repeated violations trigger maintainer review
+
+### Succession
+
+If the original author (Paul Desai) is unable to maintain the project:
+
+1. Maintainership transfers to designated successors
+2. This charter remains binding
+3. Core invariants cannot be modified by successors
+4. Community fork rights are preserved
+
+---
+
+## Amendment Process
+
+To amend this charter (except Core Invariants):
+
+1. Submit amendment RFC
+2. 60-day public comment period
+3. Security and ethics review
+4. 80% supermajority approval
+5. 6-month observation period
+6. Final ratification vote
+
+Core Invariants (Section 2) cannot be amended. They can only be added to.
+
+---
+
+## Signatures
+
+This charter is effective as of the date above.
+
+**Author:**
+Paul Desai
+Founder, N1 Intelligence
+paul@n1intelligence.com
+
+---
+
+## Changelog
+
+| Date | Version | Change |
+|------|---------|--------|
+| 2026-01-29 | 1.0 | Initial charter |
+
+---
+
+*"The purpose of governance is to make good outcomes likely and bad outcomes unlikely, even when no one is watching."*
+
+⟡ MirrorDNA — Reflection over prediction