finhay-portfolio
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a local script
./finhay.shto execute API requests, perform system health checks (doctor), and resolve user identifiers (infer). - [EXTERNAL_DOWNLOADS]: The CLI includes a
synccommand intended to update local skill definitions from an external source, functioning as a remote update mechanism. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing external data from API responses, such as order feedback messages (
feedbackmsg) and stock names, without utilizing boundary markers or sanitization logic. Mandatory Evidence Chain: - Ingestion points:
references/endpoints/order-book.md(feedbackmsg field),references/endpoints/orders.md(symbol and market_price fields). - Boundary markers: Not present in
SKILL.mdor endpoint references. - Capability inventory: Subprocess execution via
./finhay.shinSKILL.md. - Sanitization: Not specified in the instructions for handling API responses.
Audit Metadata