synthesis-voice-profiler
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill requests access to read user-provided file paths and fetch external URLs for analysis. While this is intended for processing writing samples, this capability allows the agent to access local system data if sensitive paths are provided.
- [PROMPT_INJECTION]: The skill processes untrusted content from external URLs and local files to generate its output, creating an indirect prompt injection surface.
- Ingestion points: SKILL.md Step 1 instructions allow the agent to fetch URLs and read filenames provided by the user.
- Boundary markers: None are specified in the instructions to separate untrusted sample text from the agent's internal analysis logic.
- Capability inventory: The skill utilizes file reading and network fetching capabilities.
- Sanitization: No explicit sanitization, filtering, or validation of the ingested content is described before analysis.
Audit Metadata