company-classification-routing
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to execute a local validation script,
resources/validate.py, to check the consistency of the producedclassification.jsonanddcf-result.jsonartifacts. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external financial datasets, which creates an attack surface where malicious data could influence agent reasoning.
- Ingestion points: Data is read from
mandate.json,01-data/raw-financials.json,01-data/market-data.json, and01-data/gaps.json. - Boundary markers: The skill does not define specific delimiters or instructions for the agent to ignore potentially embedded instructions in the financial data.
- Capability inventory: The skill has the capability to write diagnostic files to the filesystem and execute a local validation script.
- Sanitization: There is no description of content filtering or sanitization performed on the input data before processing.
Audit Metadata