signup

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is instructed to automatically ingest data from .agents/product-marketing.md and .claude/product-marketing.md. This represents an indirect prompt injection attack surface.
  • Ingestion points: Files located at .agents/product-marketing.md and .claude/product-marketing.md as specified in SKILL.md.
  • Boundary markers: None identified; instructions imply the agent should adopt the context directly.
  • Capability inventory: The skill does not contain executable scripts or tool invocation capabilities; behavior is limited to generating text-based recommendations and questions.
  • Sanitization: No validation or sanitization of the ingested file content is performed.
  • [NO_CODE]: The skill contains no executable scripts (e.g., Python, JavaScript, Shell) or binaries. The functionality is entirely instruction-based, delivered through Markdown files and evaluation configurations, which significantly limits the potential for direct code execution or privilege escalation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 05:32 AM
Security Audit — agent-trust-hub — signup