Vauchi Principles
The single source of truth for core principles and philosophy.
All solutions must be validated against these principles before implementation.
Core Value Statement
Vauchi is built on five interlocking commitments:
1. Privacy is a right, not an option
All design starts with: "How would we build this if users were our only concern?"
- E2E encryption for all communications
- Zero-knowledge relay (sees only encrypted blobs)
- No tracking, analytics, or telemetry
- User owns and controls their data
2. Trust is earned in person
Human recognition is the security anchor, not passwords or platforms.
- QR exchange with physical proximity verification for full trust; opt-in remote discovery at reduced trust
- No accounts or registration (device IS the identity)
- Social vouching for recovery (people you've actually met)
- No trust-on-first-use, no platform-mediated relationships
3. Quality comes from rigorous methodology
Confidence through discipline, not hope.
- Test-Driven Development (TDD) is mandatory
- Problem-first workflow with full traceability
- Threat modeling drives security decisions
- No hacks, no tech debt, no ignored tests
4. Simplicity serves the user
Vauchi respects your time and attention — it does one thing well and stays out of your way.
- No engagement tricks, no notifications designed to pull you back
- Clear, minimal interface — useful without a learning curve
- Features earn their place by solving real problems, not adding complexity
- The app is a tool, not a destination
5. Beauty adapts to the user
Simplicity and beauty go hand in hand — and beauty is personal.
- Design that feels good without demanding attention
- Theming and customisation let users make it their own
- Aesthetic choices serve clarity, never compete with it
- A beautiful tool is one that fits the person using it
Principle Categories
Privacy Principles
| Principle | Statement |
|---|---|
| User Ownership | All data stored locally on device, encrypted, under user control |
| Zero Knowledge | Relay cannot decrypt content; sees only encrypted blobs and metadata |
| No Harvesting | No analytics, telemetry, location tracking, or advertising IDs |
| No Sharing | User data never shared with third parties |
| Selective Visibility | Users control per-field, per-contact visibility |
Security Principles
| Principle | Statement |
|---|---|
| Proximity Anchors Full Trust | QR + BLE/ultrasonic required for full trust; opt-in remote contact at restricted visibility, no recovery/introduction privileges |
| Audited Crypto Only | RustCrypto audited crates primary (ed25519-dalek, x25519-dalek: Trail of Bits; sha2, hmac, hkdf: RustCrypto); chacha20poly1305 and argon2 spec-mandated; aws-lc-rs retained for TLS only (via rustls); no custom cryptography |
| Forward Secrecy | Double Ratchet ensures past messages safe if keys compromised |
| Memory Safety | Rust enforces safety; no unsafe in crypto paths |
| Defense in Depth | Multiple layers: encryption, signing, verification |
Technical Principles
| Principle | Statement |
|---|---|
| TDD Mandatory | Tidy→Red→Green→Refactor. Tidy first, test first. No exceptions |
| 90%+ Coverage | For vauchi-core; real crypto in tests (no mocking) |
| Rust Core | Memory safety, no GC, cross-platform compilation |
| Clean Dependencies | vauchi-core standalone; downstream repos use git deps |
| Gherkin Traceability | Features in features/*.feature drive test writing |
UX Principles
| Principle | Statement |
|---|---|
| Complexity Hidden | Users see "scan QR, contact added"; encryption invisible |
| In-Person Trust | Human recognition is the security anchor |
| Offline-First | QR exchange works without connectivity |
| Portable Identity | No vendor lock-in; restore from backup, switch devices |
| Cross-Platform Consistency | Same experience on iOS, Android, desktop |
Process Principles
| Principle | Statement |
|---|---|
| Problem-First | Every task starts as a problem; ideas restated as problems |
| Artifacts Accumulate | Investigation, rejected solutions, retrospectives attached to problems |
| No Wasted Rejections | Archive rejected solutions with reasoning |
| Small Atomic Commits | After each green, after each refactor |
| Retrospective Required | Learn from every completed problem |
Using These Principles
For Solution Validation
When evaluating a proposed solution, check:
- Does it align with Core Principles? (Privacy, Trust, Quality, Simplicity, Beauty)
- Does it fit the Culture? (Process Principles)
- Is it compatible with Current Implementation? (Technical Principles)
- Does it support existing Features? (UX Principles)
If a solution conflicts with any principle, it must be rejected with documented reasoning.
For Decision Making
When facing a design decision:
- Start with the user's perspective
- Assume adversarial conditions (what could go wrong?)
- Choose the option that best upholds all five core values
- Document the decision and rationale
For New Contributors
Read these principles before contributing. They are non-negotiable. If you disagree with a principle, open a problem record to discuss changing it—don't ignore it.
Amending Principles
Principles can be amended, but only through the Problem Workflow:
- Create a problem record explaining why the principle should change
- Investigate impact across codebase and documentation
- Validate the change against remaining principles
- Implement with full retrospective
Principles are not immutable, but changes must be deliberate and documented.