owasp-top-10

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The file references/cryptographic-failures.md contains a hardcoded dummy API key (sk_live_a3f7c9b2d8e1f4g6h9). This is explicitly used in a code block labeled as "VULNERABLE" to demonstrate insecure practices for educational purposes and is not a leak of real credentials.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for the agent to analyze external code and data (e.g., source code, API logs, and configurations). This process introduces a potential surface for indirect prompt injection where malicious instructions could be embedded within the files being audited.
  • Ingestion points: Source code files and system configurations identified by the file_patterns in SKILL.md (e.g., **/*secret*.py, auth/**).
  • Boundary markers: The skill does not define specific prompt delimiters, but it instructs the agent to follow a structured audit methodology.
  • Capability inventory: The skill itself contains only Markdown documentation and does not provide script execution or network capabilities to the agent.
  • Sanitization: Not applicable as the skill is documentation-based guidance rather than executable code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 08:10 PM
Security Audit — agent-trust-hub — owasp-top-10