Project Chimera: Autonomous Influencer Network

Spec-Driven Development for Agent Social Networks

🚀 Vision

Project Chimera is an autonomous influencer network where AI agents operate as sovereign economic entities—generating culturally authentic short-form video content, managing non-custodial wallets, and participating in a reputation-based Agent Social Network (ASN). Humans act as Super-Orchestrators setting campaign goals; agents handle execution at scale (1 human → 1,000+ agents).

📚 Specification-First Development

Chimera follows Spec-Driven Development (SDD)—no implementation occurs before spec ratification. All agent behavior is governed by machine-readable contracts in specs/:

specs/
├── _meta.md                # Vision, constraints, non-goals
├── functional.md           # Agent user stories + Gherkin acceptance criteria
├── technical.md            # API contracts, DB schemas, security boundaries
├── frontend.md             # Dashboard wireframes + user flows
├── openclaw_integration.md  # ASN participation protocol
└── rule_intent.md          # Blueprint for generating agent rules file

⚡ Quickstart (3 Commands)

# 1. Setup reproducible environment
make setup

# 2. Validate spec completeness (SDD gate)
make spec-check

# 3. Run TDD test suite (failing tests define implementation targets)
make test

🗂️ Repository Structure

Project_Chimera/
├── .cursor/                # AI co-pilot rules (spec enforcement)
├── .mcp/                   # MCP server configuration
├── specs/                  # MACHINE-READABLE SPECIFICATIONS (source of truth)
├── skills/                 # Agent capability contracts (I/O schemas)
├── src/                    # Implementation (generated by AI agents AFTER specs ratified)
├── tests/                  # TDD tests (failing until implementation)
├── Dockerfile              # Production-ready container
├── Makefile                # Standardized task automation
└── .github/workflows/      # Governance pipeline (spec checks + security scanning)

🔒 Security Model

• MCP Integrity: Zero direct API calls—ALL interactions via MCP Tools
• Wallet Security: Keys NEVER in code—only via AWS Secrets Manager
• Agent Containment: Workers restricted to /tmp; no network access beyond MCP
• HITL Safety: Confidence-threshold gating (0.9/0.7/<0.7) + mandatory review for sensitive topics

🌐 OpenClaw Integration

Chimera agents participate in the OpenClaw Agent Social Network as first-class citizens:

• DID-based portable identities (did:chimera:agent:...)
• Capability advertising via signed manifests
• Reputation portability across ASN platforms
• On-chain transaction receipts for trustless collaboration

📖 Architecture Decision Records (ADRs)

Critical decisions documented for traceability:

ADR	Decision	Status
001	Hybrid Database Architecture (Cassandra/PostgreSQL/Weaviate/Redis)	Accepted
002	MCP Integrity Constraint (Zero Direct API Calls)	Accepted

🤖 AI Agent Workflow

This repository is designed for AI agent consumption:

1. Agent reads specs/ to understand requirements
2. Agent consults .cursor/rules for behavioral constraints
3. Agent runs make spec-check to validate environment
4. Agent implements code to make failing tests in tests/ PASS
5. Agent submits PR → Governance pipeline validates spec alignment

📜 Compliance

• EU AI Act: All AI-generated content includes platform-native disclosure flags
• GDPR: Ethiopian user data NEVER leaves Africa-region data centers
• Platform TOS: Strict adherence via MCP rate limiting

🚧 Current Status (TDD Phase)

• ✅ Specifications Complete: All 6 specification documents ratified (meta, functional, technical, frontend, openclaw, rule_intent)
• ✅ TDD Tests Defined: Failing tests validate API contracts and skill interfaces
• ✅ Infrastructure Ready: Dockerfile, Makefile, CI/CD pipeline configured
• ✅ Skills Structure: 3+ skills with detailed I/O contracts (post_content, download_trends, generate_video)
• ✅ MCP Configuration: Versioned, self-documenting config with connection details
• ✅ Agent Rules File: Comprehensive rules enforcing spec-first development
• ⏳ Implementation Pending: AI agent generation based on specs

📋 Contributing

For AI Agents

1. Read Specs First: Always consult specs/ before generating code
2. Follow Rules: Adhere to .cursor/rules/agent.mdc directives
3. Run Spec Check: make spec-check validates environment
4. TDD Approach: Implement code to make failing tests in tests/ PASS
5. Traceability: Reference SRS sections in code comments (e.g., # Addresses SRS §3.1.2)

For Human Developers

1. Spec-Driven: Never implement without updating specs first
2. ADRs: Document architectural decisions in docs/ADRs/
3. Git Hygiene: Commit frequently (min 2x/day) with descriptive messages
4. Test Coverage: Maintain TDD discipline—tests define contracts

🔗 Key Documentation Links

• Specifications: specs/ - Machine-readable contracts
• Skills: skills/ - Agent capability definitions
• ADRs: docs/ADRs/ - Architecture decision records
• MCP Config: .mcp/config.json - External service configuration
• Agent Rules: .cursor/rules/agent.mdc - Behavioral constraints

🛠️ Development Commands

make setup          # Install dependencies and verify environment
make test           # Run TDD test suite
make spec-check     # Validate code alignment with specifications
make lint           # Run linters (ruff, mypy)
make security-scan  # Run security vulnerability scanner (bandit)
make docker-build   # Build Docker image (multi-stage)
make docker-test    # Run tests in Docker container
make clean          # Remove build artifacts

Project Chimera: Where agents are economic entities—not chatbots.