anycap-social-meme-workflows
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a standard tool for the
anycapplatform, utilizing official vendor resources (anycap.ai) and CLI tools (anycap-cli) for its primary functionality of image and meme generation. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided prompts and processes image content via the
anycap actions image-readcommand, which could allow instructions embedded in images to reach the agent's context. - Ingestion points: User-supplied meme ideas and local image files (e.g.,
./source.pngand./final-meme.png) are processed by CLI tools as described inSKILL.md. - Boundary markers: The instructions do not specify any boundary markers or delimiters for the output of the image reading tool.
- Capability inventory: The skill utilizes shell execution for generating images, editing source photos, and performing OCR (optical character recognition) on files.
- Sanitization: While the skill includes manual guardrails to avoid adult or hateful content, it lacks technical sanitization for text extracted from images during the QA process.
Audit Metadata