skill-generator

To automatically generate project-specific skills on demand, you run the skill generator to detect pattern gaps, propose new skill sets, and create SKILL.md files that extend the skill registry. It requires access to the project context and existing index.yaml.

Skill gap detection scans a project to identify missing specialized capabilities across patterns like GraphQL, CI/CD, and security. It extends the skill registry by proposing, generating, and registering new SKILL.md entries to ensure consistent orchestration without manual curation.

To add new skills to an index.yaml file, the skill generator analyzes your project context, detects patterns warranting new skills, generates the corresponding SKILL.md files, and automatically registers them within the .github/skills/index.yaml directory.

Yes, you can track deferred skill gaps in CI/CD or infrastructure projects. The generator logs deferred gaps into .github/copilot/docs/skills-opportunities.md, allowing you to monitor unaddressed patterns and re-scan the project later to discover new skills after changes.

On-demand skill generation requires an orchestrator to detect when a project lacks specialized skills. When a gap is detected, the orchestrator triggers the generator to propose skill sets, create SKILL.md entries, and update the registry accordingly.

The best way to maintain consistent skill coverage after major project changes is to re-scan the project using the skill generator. This discovers new patterns, updates the registry with generated SKILL.md files, and tracks any remaining deferred gaps in the skills-opportunities registry.