cyberelements-image2article

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • PROMPT_INJECTION (MEDIUM): Potential for Indirect Prompt Injection via vision analysis.
  • Ingestion points: The [image|url] and [path_or_url] parameters in the /cyberelements-image2article command defined in SKILL.md.
  • Boundary markers: Absent. The skill lacks instructions for the agent to ignore or isolate textual content found within images (e.g., text on screens, posters, or notes).
  • Capability inventory: The skill utilizes view_file for local file access and URL fetching for remote assets, with the ability to generate up to 2000 words of output based on these inputs.
  • Sanitization: Absent. There is no filtering or verification of the content extracted from images before it is used to generate the final article.
  • DATA_EXFILTRATION (LOW): Risk of Server-Side Request Forgery (SSRF).
  • The skill's ability to fetch and analyze images from user-supplied URLs could be exploited to probe internal network services or metadata endpoints if the agent environment is not properly network-isolated.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 12:44 PM