This project gets better when feedback is concrete.
- Add a missing PM workflow with a clear user prompt and a realistic expected output.
- Tighten an existing framework when the current guidance is too generic.
- Improve examples that help first-time users understand what "good" looks like.
- Report where the skill asks too many questions, misses a tradeoff, or fails to push back.
Include as much of this as possible:
- the prompt you used
- the output you got
- the output you expected
- the domain involved, such as discovery, roadmap, PRD, metrics, or career coaching
- whether the issue is about missing knowledge, weak reasoning, or bad interaction flow
- Keep changes scoped to one workflow or one knowledge area.
- Preserve the "push back, label assumptions, name tradeoffs" interaction style.
- Prefer improving real prompts and examples over adding generic theory.
- If you add a new workflow, add or update an example so users can see it working.
SKILL.md: routing, interaction protocol, quality gatesknowledge/: framework and domain logictemplates/: output scaffoldingexamples/: activation and trust-building demosSTARTER-PROMPTS.md: first-run prompts that should work well immediately