What problem does it solve?
Building production-grade Claude Code skills often suffers from inconsistent quality, inflated self-assessments that hide bugs, and lossy compression that breaks functionality. This skill eliminates those issues by orchestrating the full skill-building pipeline with strict, script-enforced quality gates and independent adversarial testing.
Core Features & Use Cases
- End-to-end pipeline orchestration: Automatically chains the guidance, engineer, and zipper stages with no manual handoffs, reducing human error in the skill build process.
- Anti-inflation quality verification: Uses the independent vince-attacker adversarial battery instead of self-graded checks, so skills can't fake passing quality bars.
- Bounded loop discipline: Automatically routes failures back to the owning stage (design issues to guidance, build issues to engineer, compression issues to zipper) and stops honestly if the quality bar can't be met after 3 full loops, never faking a pass.
Use case: If you have a new skill idea or an existing unpolished skill, use this to push it to industrial-grade quality without manual quality checking or risk of shipping broken skills.
Quick Start
Use the skill-conductor skill to take your new or existing skill from initial idea to production-ready, with full quality gating and independent adversarial testing.
