speak
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- Indirect Prompt Injection (SAFE): The skill ingests data from task summaries, issue reports, and plans. While this represents a surface for untrusted data, the skill provides specific 'Text Transformation' instructions that mandate the removal of URLs, code blocks, and technical IDs, effectively sanitizing the input before it is passed to TTS tools.
- Data Exposure & Exfiltration (SAFE): The skill manages state using configuration files at
.claude/tts-config.jsonand~/.claude/. Accessing these files is necessary for its stated purpose of cross-project voice management and does not involve accessing sensitive system credentials or exfiltrating data to unauthorized endpoints. - Credential Safety (SAFE): While the skill instructions reference API key identifiers (e.g., 'GOOGLE_API_KEY'), these are used as search patterns for error detection rather than hardcoded secrets. The skill correctly identifies when a provider is unconfigured and updates its state to skip it.
Audit Metadata