prompt-engineering
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely documentation-based, serving as a reference guide for designing effective prompts and interactions. It does not include executable scripts or automation tools.
- [PROMPT_INJECTION]: No malicious injection patterns or instructions to bypass safety filters were detected. The persuasion principles discussed (e.g., Authority, Commitment) are presented as legitimate strategies for improving agent instruction compliance in professional contexts.
- [EXTERNAL_DOWNLOADS]: The skill mentions the 'pdfplumber' Python package in a documentation example for text extraction, but it does not perform any actual installation or network requests.
- [COMMAND_EXECUTION]: Illustrative examples of shell commands (e.g., database migrations) and Python templates are included to demonstrate instruction specificity, but these are not executed by the skill.
- [DATA_EXFILTRATION]: The skill does not access sensitive files, environment variables, or perform network operations.
Audit Metadata