my-nanoclaw-setup-configs

I am building my team of nanoclaws. Each has its own duty/flavor (CLAUDE.md + SKILLS + MCPs)

Note: I am also leveraging the same SKILLs to train another team using Hermes AI Agent (i.e. hermes builder)

Hermes

First use idea-refine to sharpen this idea with a human-in-the-loop approach.
Treat me as an experienced collaborator, not a novice.

For the idea-refine phase:
- restate the problem clearly
- identify the target user / use case
- surface assumptions explicitly
- propose 2-4 plausible directions, not a huge brainstorm
- compare tradeoffs honestly
- recommend one direction, but pause for my validation before locking it in
- define MVP scope
- define an explicit not-doing list
- call out edge cases, failure modes, and risks early
- identify the decisions where I should weigh in before planning
- ask focused HITL checkpoints at key moments:
  1. after problem framing
  2. after option comparison
  3. before finalizing the recommended direction

Then use writing-plans to turn the validated direction into a pre-build implementation plan:
- steps
- key decisions
- edge cases
- tests/validation
- risks/tradeoffs
- likely impacted files/components

Do not code yet.

Idea:
[PASTE IDEA HERE]