speckit-red-team-run

Adversarial review of a functional specification applies trigger matching and lens selection to catch missing requirements and hidden risks before architecture lock-in. It generates a structured session report with severity-ranked findings and resolution categories.

Red teaming a spec before architecture lock-in is an adversarial analysis process that identifies ambiguous decisions and compliance gaps. It uses a catalog of adversarial lenses to target specific attack surfaces, ensuring a rigorous pre-implementation review for contractual or multi-party risks.

Generating a structured findings report for a YAML functional spec requires YAML frontmatter discovery to extract session metadata. The review produces severity-ranked findings categorized into forward-facing spec fixes, open questions, accepted risks, and out-of-scope items for audit purposes.

You should skip a red team review when trigger-aware analysis detects no relevant risk categories present in the functional specification. This mechanism avoids unnecessary work by evaluating whether the spec actually warrants adversarial scrutiny before proceeding with lens-based analysis.

Red team spec review does not modify historical working records. It distinguishes between forward-facing spec fixes and open questions while maintaining audit-friendly session metadata, ensuring past resolution tracking and working documents remain untouched.

You can override adversarial lenses during a spec review by providing explicit selections. The lens-based analysis normally selects lenses from a catalog automatically, but explicit overrides keep the review focused on the most relevant attack surfaces for your specific functional spec.