liking
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is purely informational and pedagogical. It describes psychological mechanisms and provides a framework for recognizing influence tactics without any harmful intent.
- [NO_CODE]: No executable scripts, binaries, or code files are included in the skill package. It consists solely of Markdown documentation and a JSON test case file.
- [REMOTE_CODE_EXECUTION]: There are no instances of external script downloads or execution commands (e.g., curl, wget, pip, npm).
- [DATA_EXFILTRATION]: The skill does not perform any network operations or access sensitive local files; therefore, there is no risk of data exfiltration.
- [PROMPT_INJECTION]: The instructions do not contain markers intended to bypass safety filters, override system prompts, or manipulate the underlying LLM's behavioral constraints.
Audit Metadata