tone-of-voice

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The file references/refreshing.md provides shell command examples using jq, sed, and grep. These are intended for local maintenance, allowing users to process their own message history for voice profile creation.
  • [DATA_EXFILTRATION]: The skill accesses local files in ~/.config/tone-of-voice/ to read style guidelines and message excerpts. These files are kept local and are not transmitted over the network. The skill includes explicit guidance on redacting sensitive personal information from these files.
  • [PROMPT_INJECTION]: The skill processes user-supplied voice profiles, which technically constitutes an indirect prompt injection surface. However, the skill provides a comprehensive 'Final self-check' and a list of 'banned words' and 'AI tells' to ensure the output remains safe and authentic to the user's voice.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 08:04 AM
Security Audit — agent-trust-hub — tone-of-voice