Skills
skills/samber/cc-skills/copywriting-tone-of-voice-creator/Gen Agent Trust Hub

copywriting-tone-of-voice-creator

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating untrusted data into instructions for sub-agents and generated documentation.\n
  • Ingestion points: Untrusted data enters the context through WebFetch during Phase 2 (market research) and by reading the local SOUL.md file in Phase 1. User-supplied variables like <category> are also directly interpolated into sub-agent briefs.\n
  • Boundary markers: The skill does not employ explicit delimiters or "ignore embedded instructions" warnings when processing these untrusted inputs.\n
  • Capability inventory: The skill possesses significant capabilities, including Write, Edit, and the Agent tool for spawning sub-tasks with their own instructions.\n
  • Sanitization: No sanitization, escaping, or validation is performed on external content or user-supplied strings before they are interpolated into prompts for the research sub-agent or written to the TONE.md file.\n- [EXTERNAL_DOWNLOADS]: The skill utilizes WebSearch and WebFetch to access various non-whitelisted external domains during Phase 2 research. While intended for brand analysis, this facilitates connections to arbitrary remote sources.\n- [COMMAND_EXECUTION]: The skill performs routine file system management using Read, Write, Glob, and Edit tools to create and modify brand documentation files such as TONE.md and SOUL.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 12:49 AM
Security Audit — agent-trust-hub — copywriting-tone-of-voice-creator