article-enrichment

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process raw article text dumps, which are external and untrusted data sources. This creates a surface for indirect prompt injection attacks where malicious instructions embedded in an article could attempt to influence the agent's behavior during the enrichment phase.
  • Ingestion points: The agent reads raw content from article pages via gbrain get media/articles/<slug> as described in the 'Invocation' section of SKILL.md.
  • Boundary markers: The skill instructs the agent to wrap the original raw source in a <details> block, but the LLM must process the raw text directly to generate the summary and insights, providing no inherent boundary during the transformation phase.
  • Capability inventory: The skill has the capability to write files back to the filesystem under media/articles/ using the gbrain put_page operation.
  • Sanitization: There is no explicit sanitization or filtering of the raw article content mentioned before it is processed by the LLM (Sonnet or Opus).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 06:18 PM
Security Audit — agent-trust-hub — article-enrichment