brain-ingest-gate

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 untrusted external content for migration into a central repository. Malicious instructions embedded in this source material could potentially influence the agent during critical steps such as core claim extraction, named-entity resolution, or alias generation, leading to metadata poisoning or incorrect indexing.
  • Ingestion points: External files, bulk imports, and workspace content targeted for migration via triggers like 'move this to brain'.
  • Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when processing untrusted content.
  • Capability inventory: Employs platform-specific tools (gbrain) to read from and write to the filesystem across various directories (people/, companies/, concepts/, projects/).
  • Sanitization: There is no mention of sanitizing or validating user-derived strings before they are interpolated into search queries or entity resolution commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 06:18 PM
Security Audit — agent-trust-hub — brain-ingest-gate