molten-brand

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns detected. The skill instructions are limited to gathering user input via conversational chat or structured question tools and writing a markdown file to a local path (molten-docs/brand/brand.md).
  • [COMMAND_EXECUTION]: The skill uses standard agent platform tools for user interaction (e.g., AskUserQuestion, request_user_input). These are used for legitimate interaction and do not involve shell command execution or privilege escalation.
  • [PROMPT_INJECTION]: No evidence of prompt injection or attempts to bypass AI safety guidelines. The instructions reinforce safety by explicitly telling the agent not to generate application code or visual design details.
  • [DATA_EXFILTRATION]: No network operations or external URLs were found. The skill operates entirely within the local repository context.
  • [EXTERNAL_DOWNLOADS]: No external dependencies, packages, or remote scripts are referenced or required for this skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 02:47 PM
Security Audit — agent-trust-hub — molten-brand