general-ABA-keyword-monitor
Pass
Audited by Gen Agent Trust Hub on Apr 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill manages its operational environment by executing
pip installvia a subprocess inconfig.pyto ensure necessary Python libraries are available. - [COMMAND_EXECUTION]: Automates report generation by invoking the
lark-cliutility through subprocess calls inbitable_reporter.py, enabling the creation of multi-dimensional tables and dashboards in the user's Feishu environment. - [EXTERNAL_DOWNLOADS]: Fetches Amazon search data from
amz123.comand historical market trends from the Sorftime service (mcp.sorftime.com), which are well-known resources in the e-commerce industry. - [PROMPT_INJECTION]: Features an indirect prompt injection surface where external data scraped from the web is processed by an LLM for classification in
classifier.pyand analysis inmain.py. This is documented as a functional feature for data interpretation, and the risk is mitigated by the structured nature of the tasks.
Audit Metadata