Thank you for your interest in contributing to the Industrial Agentic Intelligence Framework.
- Read the Code of Conduct.
- Open an issue to discuss significant changes before submitting a PR.
- For security vulnerabilities, follow the process in SECURITY.md.
git clone https://github.com/adris-misra/multi-agentic-framework.git
cd multi-agentic-framework
make install- Fork the repo and create a branch:
git checkout -b phase-N/short-description(orfeat/short-descriptionfor community contributions). - Write code + tests. Run
make lint typecheck testbefore pushing. - Open a PR against
main. Use the PR template in.github/PULL_REQUEST_TEMPLATE.md.
- Formatter/linter:
ruff(configured inpyproject.toml; runmake format). - Type checker:
mypy --strictonsrc/(runmake typecheck). - Line length: 100 characters.
- Docstrings: Google style.
- No
print()insrc/; usestructlog.
- Unit tests:
tests/unit/— no I/O, no LLM calls, no docker. - Integration tests:
tests/integration/— require the docker-compose stack (make demo). - Eval tests:
tests/eval/— LLM-quality regressions; nightly only. - Target ≥ 80% line coverage on
src/.
Follow Conventional Commits:
feat(agents): add bearing-wear anomaly detection heuristic
fix(tools): handle OPC UA reconnect on timeout
docs(standards): add ISA-88 batch recipe alignment notes
- All UNS paths must be validated against
config/uns_topic_taxonomy.yaml. - Any agent action that mutates state at Purdue zone ≤ 2 must pass Safety/Guardrail review.
- Do not add LLM providers beyond those in §2.3 of the spec without a prior issue.