investor-panel
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted financial data to drive the reasoning and output of simulated investor personas, creating a vulnerability surface where instructions embedded in the data could influence agent behavior.
- Ingestion points: The skill reads stock metrics and raw data from
.cache/{ticker}/dimensions.jsonand.cache/{ticker}/raw_data.jsonas described in the execution steps ofSKILL.md. - Boundary markers: The instructions do not define any delimiters or provide "ignore embedded instructions" warnings to the model before it processes the external data files.
- Capability inventory: The skill allows the agent to generate complex reasoning and formatted JSON signals which are persisted to
.cache/{ticker}/panel.json. - Sanitization: There is no logic provided to sanitize or validate the content of the ingested JSON files to ensure they do not contain natural language instructions that could override the primary prompt.
Audit Metadata