Clément SAILLANT
17c3ad6c78
- Introduced PM Agent and QA Agent with detailed plans and objectives. - Created prompts for generating intake, specifications, plans, and tasks. - Developed templates for architect, doc, firmware, and hardware schematic agents. - Added installation and coordination plans to streamline multi-agent workflows. - Established a comprehensive mapping and recommendations for the Kill_LIFE project. - Included a flow diagram and technical analysis for better understanding of the project structure.
1.1 KiB
1.1 KiB
Analyse technique Kill_LIFE
Scripts clés
- scope_guard.py : enforcement CI, mapping labels → allowlist/denylist, sécurité PR.
- cockpit.py : orchestration agents/gates, automatisation des tâches.
- compliance/validate.py : validation des profils, mapping exigences/evidence.
- hw/schops : bulk edits, exports, checks hardware.
- ai/compose_codex_prompt.py : génération de prompts pour Codex.
Evidence pack
- Généré à chaque gate, mappé dans compliance/plan.yaml.
- Traçabilité des artefacts, logs, rapports.
Tests
- firmware/test/ : tests unitaires Unity.
- hardware : ERC, netlist, BOM, DRC.
- CI/CD : validation automatique, enforcement des gates.
Sécurité
- sanitizer : nettoyage des issues/PR.
- OpenClaw : observateur, actions auditées.
Automatisation
- bulk edits hardware, orchestration agents, validation compliance.
Points de friction
- Onboarding dense, passage de gates, gestion des exceptions.
Robustesse
- Bonne couverture CI/CD, sécurité renforcée, traçabilité exemplaire.
Pour chaque script, prévoir des tests unitaires, une documentation d’usage, et une intégration CI.