steering-specs-generator

Tacit knowledge capture uses structured interviews to surface undocumented rules, conventions, and practices. This Skill guides stakeholders through context-aware interviews, extracting engineering knowledge and generating standardized steerings in Intent → Rules → Practices → Meta format that agents can follow and enforce.

Generate steerings by configuring output paths, defining topics, and running parallel discovery across repository context, code conventions, architecture invariants, and business domain contracts. The Skill spawns sequential pack-interview agents that produce steerings documents and backlog items in your configured directories.

Yes. The Skill supports both predefined packs and custom topics, letting you encode team-specific conventions for quality, style, testing, security, compliance, and delivery lifecycle. Each interview generates structured steerings applicable to your codebase topology and business requirements.

The interview workflow covers risk history, security, and compliance as built-in discovery dimensions alongside code topology and quality standards. Interview agents extract these concerns and encode them into steerings that provide context for decision-making and compliance verification.

The Skill outputs steering documents in Intent → Rules → Practices → Meta format and generates corresponding backlog items. Steerings are written to a configured steerings directory and backlog items to a backlog directory, both ready for agent consumption and team reference.