annual-revenue
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill navigates to and extracts text content from an external website (Yahoo Finance) using Chrome DevTools.
- Ingestion points: Data is ingested from the DOM of
finance.yahoo.comvia theevaluate_scripttool in Step 3, specifically targeting table headers and row titles/values. - Boundary markers: The skill lacks explicit boundary markers or instructions for the agent to ignore potentially malicious text embedded within the financial data it scrapes.
- Capability inventory: The skill has access to the
Bashtool,Grep, and theTasktool (which allows it to spawn and manage subagents). - Sanitization: There is no sanitization or filtering of the extracted text content before it is processed by the agent to build the final markdown tables.
- [DYNAMIC_EXECUTION]: The skill uses the
evaluate_scripttool to execute JavaScript within the browser context. However, the script provided is static and intended solely for data extraction, minimizing the risk of arbitrary code execution.
Audit Metadata