market-audit
Pass
Audited by Gen Agent Trust Hub on Apr 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it consumes untrusted data from arbitrary URLs.\n
- Ingestion points: The
WebFetchtool is used in Phase 1.1 to retrieve content from a user-provided target URL and key interior pages which are outside the control of the agent environment.\n - Boundary markers: The prompt does not define clear delimiters (such as XML tags or Markdown code blocks with headers) to wrap the fetched content, nor does it instruct the subagents to disregard or ignore any instructions found within the downloaded site data.\n
- Capability inventory: The skill has the ability to write a detailed report file (
MARKETING-AUDIT.md), display data in the terminal, and initiate five parallel subagent processes, making it a powerful orchestration point for potentially poisoned data.\n - Sanitization: There is no evidence of sanitization, filtering, or validation of the HTML or text content retrieved from the web before it is passed to the subagents for analysis.
Audit Metadata