guideline-generation

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external data, including brand documents and meeting transcripts, which could contain malicious instructions designed to influence the agent's synthesis or quality assurance phases.
  • Ingestion points: Document files and transcripts identified in SKILL.md.
  • Boundary markers: None explicitly mentioned in the processing logic.
  • Capability inventory: File system writes and archival operations in SKILL.md, and delegation to specialized agents.
  • Sanitization: Instructions are provided to redact customer names and anonymize sensitive information.
  • [COMMAND_EXECUTION]: The skill performs file system management within the user's working directory, specifically reading from .claude/brand-voice.local.md, writing to .claude/brand-voice-guidelines.md, and archiving older versions via rename operations. The instructions include security-conscious logic to resolve paths relative to the user's working folder to prevent unauthorized file access.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 03:21 PM
Security Audit — agent-trust-hub — guideline-generation