poster-information-hierarchy
Pass
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill contains numerous references to external media assets and portfolio sites in
references/cases.md(e.g.,goodcase.ai,x.com,liblib.art,runninghub.ai). The instructions inSKILL.mdexplicitly direct the agent to inspect these remote images using available browser or image tools to ensure visual fidelity during the prompt generation process. - [INDIRECT_PROMPT_INJECTION]: The skill processes structured data from
references/cases.md, which includes summaries and prompt excerpts from various third-party creators. The workflow mitigates risks by requiring the agent to generate a 'reference contract' (Preserve, Replace, Avoid) that explicitly identifies which structural traits are being adopted, ensuring the external content is used as a template rather than an unvetted instruction source.
Audit Metadata