conversation-qa

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by capturing and persisting conversation content into a generated document.\n
  • Ingestion points: User-provided questions and dialogue turns are extracted from the conversation history (SKILL.md Step 1).\n
  • Boundary markers: The skill lacks instructions to use boundary markers or to ignore instructions that might be embedded within the conversation text.\n
  • Capability inventory: The skill uses Write and Edit tools to create and update local HTML documents, and Bash to execute ls -t for file management (SKILL.md frontmatter and Step 2).\n
  • Sanitization: There are no explicit instructions to escape HTML entities or sanitize the conversation data, presenting a potential Cross-Site Scripting (XSS) risk in the generated document.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 12:57 AM
Security Audit — agent-trust-hub — conversation-qa