verify

You can verify BDD scenarios by dropping .feature.md files into the bdds/ directory, which the agent ingests to parse YAML frontmatter, plan verification, execute steps via the interaction surface, and collect behavioral evidence.

Black-box BDD verification confirms that written scenarios faithfully reflect runtime system behavior by observing external interactions only, reducing implementation bias and drift without requiring access to internal code details.

Regression checks across related BDD areas involve ingesting scenarios, planning verification, executing steps through the configured interaction surface, and updating the knowledge graph to detect regressions and ensure system stability.

Yes, BDD verification updates a knowledge graph by ingesting scenarios, parsing YAML frontmatter, executing steps via the interaction surface, and collecting evidence according to the configured evidence level to map behavioral correctness.

Yes, during QA verification sessions, the system ingests new .feature.md files arriving in bdds/ and parses their YAML frontmatter to extract area, tags, and notes for planning and executing behavioral verification steps.

Black-box BDD verification operates purely against observable behavior and does not peer into implementation details, meaning it cannot diagnose internal code-level defects directly and relies entirely on the configured interaction surface and evidence collection levels.