classify-post-to-section

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from the body of blog posts to perform classification tasks. This creates a surface for potential indirect prompt injection where malicious text within a post could attempt to override the classification logic.
  • Ingestion points: The agent is instructed to read the full body of a post (SKILL.md).
  • Boundary markers: The instructions lack explicit boundary markers or specific guidance to ignore instructions embedded within the post content.
  • Capability inventory: The skill reads local configuration files (section-map.md) and returns structured classification metadata.
  • Sanitization: No sanitization, validation, or filtering of the post content is mentioned in the workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 02:36 PM
Security Audit — agent-trust-hub — classify-post-to-section