generate-creative-ideas
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No malicious bypasses or role-play injections (DAN-style) were detected. Phrases such as "ignore existing conventions" in the Tabula Rasa template are clearly scoped to creative brainstorming exercises and do not attempt to override agent system constraints.
- [Remote Code Execution] (SAFE): The skill contains no scripts, binaries, or command-line instructions. It is a data-only skill providing prompts and methodologies.
- [Data Exposure & Exfiltration] (SAFE): No sensitive file paths, environment variables, or hardcoded credentials were found. All external links point to reputable academic and professional sources (Nature, HBR, arXiv, etc.) for research purposes.
- [Indirect Prompt Injection] (LOW): The skill provides templates that interpolate user input (e.g.,
[PROBLEM],[GOAL]). While this creates a surface for indirect injection if the user provides malicious content, the lack of tool execution or file-system capabilities in the skill prevents any significant impact. The presence of these templates is essential to the skill's primary purpose. - [Obfuscation] (SAFE): Content is clearly written in plain markdown. No hidden characters, Base64 encoding, or homoglyph attacks were identified.
Audit Metadata