brand-guidelines

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions define professional guidelines for brand voice and visual identity. No malicious patterns, obfuscation, or unauthorized data access were detected.
  • [PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by reading from a local file, though this is considered safe due to lack of capabilities.
  • Ingestion points: Reads .marketing-os/product-context.md in SKILL.md.
  • Boundary markers: Absent.
  • Capability inventory: None detected; the skill does not use subprocesses, network tools, or file-writing operations.
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 12:11 AM
Security Audit — agent-trust-hub — brand-guidelines