skills/jeremylongshore/claude-code-plugins-plus-skills/evaluating-machine-learning-models/Gen Agent Trust Hub
evaluating-machine-learning-models
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through the processing of external data files. * Ingestion points: Data is ingested via
scripts/data_loader.py(JSON loading) andassets/visualization_script.py(CSV loading via pandas). * Boundary markers: No boundary markers or instructions to ignore instructions within data files are implemented in the skill definition. * Capability inventory: The skill allows extensive tool usage includingBash(cmd:*),Write,Edit,Grep, andGlob. * Sanitization: No sanitization or validation of the ingested data is performed before it is processed by the agent. - [COMMAND_EXECUTION]: The skill definition requests broad system permissions via
allowed-tools: Bash(cmd:*), which allows for arbitrary command execution. While intended for model evaluation, this capability increases the potential impact of indirect prompt injection if malicious instructions are present in processed datasets.
Audit Metadata