doc-bdd-autopilot
Pass
Audited by Gen Agent Trust Hub on Mar 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill leverages local automation by invoking Python scripts and shell hooks within the 'ai_dev_ssd_flow' directory. These executions are dedicated to validating documents and generating BDD suites, which aligns with the skill's stated purpose.
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it ingests untrusted requirements from 'docs/03_EARS/' to generate BDD scenarios. Ingestion points: EARS markdown files. Boundary markers: Not explicitly defined. Capability inventory: File writing and local script execution (python/bash). Sanitization: Performs functional syntax validation but lacks specific security-focused sanitization.
- [SAFE]: All operations are confined to local project directories. No unauthorized network calls or credential exposures were identified.
Audit Metadata