The knowledge base for building reliable AI agents — context engineering, orchestration, agent design, and the Research → Plan → Implement process. Hands off to OpenBuild when it's time to build.
superbuild v1.1 serves three roles: a knowledge repository for agentic engineering, an action framework (RPI: Research → Plan → Implement), and a skill development guide (build/test/debug/package/distribute). It covers 5 design patterns and provides targeted reference files for each common situation. When done, it produces RESEARCH.md + PLAN.md for handoff to OpenBuild.
Knowledge repository — read only what you need:
| Situation | Reference file |
|---|---|
| Context window filling up | context-engineering.md |
| Multi-agent coordination | orchestration.md |
| Spawn sub-agents, pick models | agent-design.md |
| Memory organization | memory-patterns.md |
| Reduce tool call overhead | tool-optimization.md |
| Quick tips and lessons | tips.md |
RPI Process: Research → Plan → Implement Skill guide: build, test, debug, package, distribute
Start with the situation table and read only the matching reference file. Handoff to OpenBuild: provide RESEARCH.md + PLAN.md.
superbuild, RPI, agentic, agent design, workflow design, context engineering, build skill, test skill, publish skill, skill not working
- openbuild — project lifecycle (receives RESEARCH.md + PLAN.md)
Part of the NachaFromMars OpenClaw skill ecosystem.