Project Chimera is a spec-driven repository that scaffolds an autonomous AI influencer platform: trend ingestion, content generation, safety reviews, and service integration (OpenClaw). This repo contains the technical and functional specifications, schema contracts, skill definitions, test harnesses, CI/CD workflows, and infrastructure scaffolding needed to develop, validate, and operate the agents.
Project Chimera is designed to provide a production-ready, competition-level framework for autonomous AI agents. It ingests social trends, generates content variants, enforces skill contracts, validates against schemas, and maintains observability and governance through CI/CD pipelines.
project-chimera-agent-factory/
├── README.md
├── LICENSE
├── specs/
│ ├── _meta.md
│ ├── functional.md
│ ├── technical.md
│ └── openclaw_integration.md
├── skills/
│ ├── skill_download_youtube/
│ │ └── HANDLER_PLACEHOLDER.md
│ ├── skill_transcribe_audio/
│ │ └── HANDLER_PLACEHOLDER.md
├── tests/
│ ├── conftest.py
│ ├── unit/
│ └── integration/
├── scripts/
│ └── validate_skill_contracts.py
├── alembic/
├── Dockerfile
├── Makefile
└── .github/
└── workflows/
- AI/ML Engineers building autonomous agent systems.
- DevOps engineers managing CI/CD and containerized environments.
- Researchers exploring spec-driven agent architectures.
- Teams aiming for competition-level, production-ready Python projects.
- Python 3.11+
- Docker (optional but recommended)
- Git
- Basic understanding of:
- OpenAPI
- JSON Schema
- Test-driven development
Clone the repository:
git clone https://github.com/hydropython/project-chimera-agent-factory.git
cd project-chimera-agent-factoryThis project uses uv for fast, reproducible Python environments:
uv venv
source .venv/bin/activate # Windows: .\.venv\Scripts\activate
uv pip install -r requirements.txtmake spec-checkValidates:
- OpenAPI contracts
- JSON schemas
- Spec integrity
python src/validate_skill_contracts.pyEnsures:
- Skill inputs/outputs match schemas
- No undocumented interfaces exist
make test- JSON inputs/outputs as defined in
specs/schemas. - Video metadata stored in Postgres (see
specs/db_schema.md). - Skills input/output must conform to JSON schema contracts.
- Tests are written before implementation (True TDD).
- Failing tests represent "empty slots" that agents must implement.
- Agents must conform to tests, not the other way around.
- CI/CD configuration:
.github/workflows/main.yml - AI co-pilot rules:
CLAUDE.md - AI review policy:
.coderabbit.yaml - Environment variables:
.env(not committed)
Project Chimera follows these principles:
- Specification First
- Schemas as Law
- Tests as Contracts
- Skills as Capabilities
- Governance over Freedom
No code is written without specs.
No specs exist without validation.
- Schema validation: milliseconds
- Skill contract checks: milliseconds
- CI execution: < 5 minutes
- Docker image: minimal, deterministic
Build the Docker image:
make build-imageRun inside Docker:
docker run chimera-agent-factory- CI runs on every push and PR:
- Spec validation
- Skill contract validation
- Test execution
- AI governance rules enforce:
- Spec alignment
- Security hygiene
- No undocumented behavior
- Secrets are never committed.
- Key management via external KMS is recommended.
.coderabbit.yamlchecks for insecure patterns.
- Read
CLAUDE.md(Prime Directive) - Reference specs in every PR
- Add or update tests
- Keep commits small and reviewable
- PRs without spec alignment will be rejected
See the top-level LICENSE file for full license terms.
See Git commit history for versioned changes.
If you use this project in academic work, cite via:
Project Chimera — Hydropython Team, 2026. Repository: https://github.com/hydropython/project-chimera-agent-factory
- Repository owner: see Git history and remote
origin - Issues: open GitHub issues in this repository
flowchart LR
subgraph External
A["Social Platforms\n(Twitter, Reddit, YouTube, Google Trends)"]
B["OpenClaw Network"]
H["Human Reviewers"]
end
subgraph Chimera_Core
TF["Trend Fetcher"]
TA["Trend Analyzer"]
CG["Content Generator"]
SL["Safety Layer"]
SK["Skills Runtime"]
RG["Registry / Integration"]
DB["Metadata DB"]
CACHE["Cache / Vector DB"]
OBS["Observability / Tracing"]
end
A -->|signals| TF
TF --> TA
TA --> DB
TA --> CG
CG --> SL
SL -->|queue| H
H -->|decision: approve or reject| SL
SL -->|publish| RG
RG -->|register or discover| B
RG --> SK
SK --> CG
CG --> DB
DB --> CACHE
TF --> OBS
TA --> OBS
CG --> OBS
RG --> OBS