standards-compliance-training
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHPROMPT_INJECTION
Full Analysis
- PROMPT_INJECTION (HIGH): The skill is designed to ingest and process external content ('training programs', 'regulatory requirements', 'content') to verify compliance. This constitutes a significant Indirect Prompt Injection surface.
- Ingestion points: The 'content' parameter in
skill.pyand various CLI flags like--training,--program, and--contentinSKILL.mdaccept external data. - Boundary markers: None detected. There are no instructions to the agent to disregard embedded commands in the training materials it analyzes.
- Capability inventory: While
skill.pyis currently a stub, theSKILL.mddocumentation describes complex file system operations ('safety-program/', 'hospital-orientation/'). - Sanitization: None detected in the provided code.
- COMMAND_EXECUTION (MEDIUM): The
SKILL.mddefines a CLI interface that implies the ability to execute file-system level audits and report generation, which may lead to unauthorized file access if not properly scoped. - NO_CODE (LOW): The
skill.pyfile contains undefined variables (skill_dir) that are not initialized or imported, which will result in aNameErrorat runtime. This indicates the skill is either incomplete or poorly tested.
Recommendations
- AI detected serious security threats
Audit Metadata