india-stock-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill combines ingestion of untrusted data from web searches with access to tools that can read sensitive user financial data, which is a classic indirect prompt injection pattern.
- Ingestion points: The skill instructs the agent to use web search to gather news, analyst reports, and sector developments for stock analysis in Analysis Types 1, 2, and 4 in SKILL.md.
- Boundary markers: The instructions do not include prompt boundary markers or directives for the agent to disregard instructions found within the external data retrieved from web searches.
- Capability inventory: The skill utilizes Zerodha Kite MCP tools such as get_holdings, get_positions, and get_profile which allow the agent to read the user's private financial data and personal profile.
- Sanitization: No sanitization or validation mechanisms are mentioned to mitigate the risks of processing untrusted content from the web.
Audit Metadata