article-enrichment

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes raw article text dumps which are considered untrusted data and could contain embedded instructions designed to influence the agent's behavior.
  • Ingestion points: The skill reads raw content from article files located in media/articles/ using the gbrain get command (defined in SKILL.md).
  • Boundary markers: While the skill instructs the agent to wrap the original raw source in a <details> block in the output, it does not explicitly instruct the agent to ignore instructions that might be contained within that raw text during the enrichment process.
  • Capability inventory: The skill has the ability to read files, query the brain via gbrain, and write new structured pages to media/articles/ using put_page.
  • Sanitization: There is no evidence of filtering, escaping, or validation of the article content to remove potential prompt injection vectors before the agent processes the text.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 04:37 PM
Security Audit — agent-trust-hub — article-enrichment