writing

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches product documentation from the official Sablier domain (docs.sablier.com/llms.txt) to maintain technical accuracy during content generation. This is a trusted vendor resource used for contextual information.
  • [DATA_EXFILTRATION]: No unauthorized data collection or external transmission of sensitive information was identified. The skill instructions restrict data access to local brand references and official documentation.
  • [PROMPT_INJECTION]: The instructions are focused on content creation and brand alignment. No patterns were found attempting to override safety guidelines or bypass agent constraints.
  • [COMMAND_EXECUTION]: The skill contains no shell commands, script execution patterns, or dynamic context injection via the !command syntax.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 12:59 AM
Security Audit — agent-trust-hub — writing