brand-voice-extractor
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content fetched from external web pages, which is an untrusted data source. This creates a surface for indirect prompt injection where the agent might interpret text from those pages as instructions rather than data.
- Ingestion points: The skill uses
WebFetchin Phase 2 to retrieve text from 10-20 external URLs. - Boundary markers: The instructions do not define any delimiters or provide a 'system-style' warning to the agent to ignore any instructions embedded within the fetched text.
- Capability inventory: The skill utilizes
WebFetch, file read (content inventory), and file write (profile generation) capabilities. - Sanitization: There is no logic for sanitizing or filtering the content before it is passed to the analysis phase.
- [EXTERNAL_DOWNLOADS]: The skill is primarily built around fetching content from external, user-specified or cataloged domains.
- [NO_CODE]: This is an agent-executed skill with no included scripts, meaning it relies entirely on the underlying AI model's instructions and tool-calling capabilities.
Audit Metadata