mint-report
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
bashto search the local filesystem for financial records such as payroll and invoices. This behavior is consistent with the stated purpose of financial reporting. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from local CSV and Markdown files without sanitization or boundary markers.
- Ingestion points: Step 0 searches local files (CSV, MD) for financial keywords (revenue, payroll, invoices) to ingest.
- Boundary markers: No delimiters or warnings are used to separate ingested file content from the agent's instructions.
- Capability inventory: The skill utilizes
BashandReadtools to access and process the data. - Sanitization: No sanitization or validation of the ingested file content is performed.
Audit Metadata