editorial-portrait-composition
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an Indirect Prompt Injection surface. It instructs the agent to read and inspect external content from
goodcase.ai,youmind.com, andx.combefore drafting responses. - Ingestion points: The
references/cases.mdfile contains numerous external links to 'GoodCase evidence' and 'finished media' that the agent is explicitly told to 'Inspect' and 'Read'. - Boundary markers: The skill lacks explicit boundary markers or instructions to disregard potential commands found within the external evidence files.
- Capability inventory: The skill utilizes image/browser tools to inspect remote media files and generates complex prompt artifacts for downstream AI image generators.
- Sanitization: No evidence of sanitization or filtering of the content retrieved from external URLs was found in the instructions.
- [SAFE]: The skill uses external links to well-known social media platforms (X/Twitter) and specialized creative portfolio sites (GoodCase) for reference purposes, which is standard for an image prompting tool. These are documented neutrally as they align with the skill's primary purpose of providing editorial reference material.
Audit Metadata