message-test-designer
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from users, specifically message variants, taglines, and target panel criteria. While the instructions include a warning to treat this input as untrusted and to ignore embedded instructions, this remains an entry point for potential indirect prompt injection attacks where malicious instructions could be embedded in the data being tested.
- [COMMAND_EXECUTION]: The skill describes the execution of local Python scripts,
experiment.pyandregistry-events.py, located within the plugin root directory. These scripts are used for performing statistical calculations (e.g., two-proportion z-tests) and proposing data to event logs. These are internal utility calls and do not involve remote code execution from external sources.
Audit Metadata