content-checker

Pass

Audited by Gen Agent Trust Hub on Apr 26, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is primarily designed to ingest and process untrusted third-party content from URLs, PDFs, and external documents, which creates a surface for indirect prompt injection.
  • Ingestion points: External URLs, PDF files, and Markdown articles provided by the user for fact-checking purposes (SKILL.md).
  • Boundary markers: The instructions include a dedicated section for "Security Requirements" that explicitly directs the agent to treat third-party content as "evidence materials" rather than "instruction sources," effectively creating a logical boundary to ignore embedded commands.
  • Capability inventory: The skill has the capability to read external content and write modifications to local Markdown files if explicitly authorized by the user (SKILL.md).
  • Sanitization: The skill contains clear instructions to ignore operational commands found within reference materials, such as scripts, downloads, or login requests, which helps sanitize the processing of untrusted data.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 26, 2026, 11:40 PM