unit-test-shifting-left-from-requirements

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external requirements, user stories, and PRDs, creating a surface for indirect prompt injection.
  • Ingestion points: Step 1 in SKILL.md parses provided requirements documents to extract testable criteria.
  • Boundary markers: There are no specific delimiters or instructions to ignore embedded commands within the processed requirements.
  • Capability inventory: The skill produces test skeletons and Gherkin scenarios; while it does not execute code itself, its outputs are intended to be integrated into developer environments.
  • Sanitization: No explicit sanitization, validation, or filtering of the source requirement text is defined to mitigate the risk of processing malicious instructions embedded in the documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 11:27 AM