senpi-why
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes a workflow where the agent ingests a "thesis in plain English" from a user (e.g., market predictions) and automatically "builds and runs" a corresponding trading strategy. This documentation highlights a potential surface for indirect prompt injection if the underlying execution engine does not properly delimit or sanitize user-supplied natural language before translating it into financial actions.
- Ingestion points: User-provided market theses provided in plain English (described in SKILL.md).
- Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are mentioned for the strategy-building phase in these instructions.
- Capability inventory: The documentation mentions tools for sizing, exiting positions, and running strategies (e.g.,
get_loyalty_tiers,market_list_instruments), implying significant financial execution capabilities. - Sanitization: Not present in the analyzed instructional files.
Audit Metadata