
Honest, criteria-based.
Your own prompts know your context; the library knows the failure modes. The answer is both — and we built the import path.
| Criterion | PM Skills | The alternative |
|---|---|---|
| Startup cost | Zero: 526 skills now | Hours per prompt, compounding maintenance |
| Failure modes | 2,500+ anti-patterns baked in from practice | You discover them one bad deliverable at a time |
| Your context | Bring it: CONTEXT.md, the Brain, Skill Remix, `import cursor|cline` converts your existing rules | Fully yours by definition |
| Team consistency | Same skill = same structure for everyone; org server for private libraries | Every teammate's prompts drift apart |