brand-voice-profiler
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were identified in the skill instructions or reference files. The skill's behavior is consistent with its stated purpose of brand analysis.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill performs file read operations on './brand/brand-dna.md' and write operations to './brand/brand-voice.md'. These actions are confined to the project directory and are consistent with the skill's stated purpose of creating brand documentation.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources by scraping website content provided by the user. While this introduces a surface for indirect prompt injection, the risk is mitigated by the structured process and template-based output requirements defined in the skill. (1) Ingestion points: Website URLs provided by the user for scraping via WebFetch or Firecrawl (SKILL.md, Step 2). (2) Boundary markers: None explicitly defined in the prompt instructions for the scraping tool output. (3) Capability inventory: Reads './brand/brand-dna.md', reads 'references/' files, and writes to './brand/brand-voice.md' (SKILL.md, Steps 1 & 7). (4) Sanitization: Not explicitly defined; the agent is instructed to extract specific patterns from the text.
Audit Metadata