Architecture Decision Records

Design decisions that shape the Society.

ADR-001: Markdown-Based Skills Over Code-Based Agents
ADR-002: Conventional Commits as the Commit Standard
ADR-003: Dual Enforcement via Bash Hooks and Commitlint
ADR-004: 16 Specialized Members Over a General-Purpose Agent
ADR-005: Orchestrator Pattern Over Peer-to-Peer Member Communication
ADR-008: TypeScript-Native Runtime Over Python + DeepAgents
ADR-009: Groq as the Default Free LLM Provider
ADR-010: LanceDB for Vector Storage
ADR-011: Rate Limiter and State Store
ADR-012: Error Handling and Resilience Strategy
ADR-013: Distribution Channel Priority
ADR-014: EvalResult in Core Types
ADR-015: Decision Intelligence and Provenance
ADR-016: LLM-as-Judge Evaluation Pipeline
ADR-017: Semantic Learning via Embedding Index
ADR-018: Redaction Scope Extended to Decision and Provenance Payloads
ADR-019: Executor contract returns model attribution
ADR-020: Mind Virus Mitigation Strategy
ADR-021: Society Projection as a Live Surface
ADR-022: Monetisation Model
ADR-023: Third-Party Member Plugin API
ADR-024: Tiered Memory Model with Per-Tier Bounds
ADR-025: ReAct Tool Loop for Member and Core-Agent Runs
ADR-026: tree-sitter for Structural Code Analysis
ADR-027: Provider Chain with Circuit Breaker and Per-Request Timeout
ADR-028: SKILL.md Frontmatter as an Enforced Contract
ADR-029: Fail-Closed Skill Trust and Prompt-Injection Detection
ADR-030: Metrics Export Schema and Privacy Boundary