voice
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a standard local-only utility for writing style analysis. It reads and writes to project-specific configuration files (
social-context.md) within the agent's intended environment. - [DATA_EXPOSURE]: The skill processes user writing samples, including potentially sensitive emails. However, the data processing is restricted to stylistic analysis (counting sentence length, punctuation, vocabulary) and remains within the local context. No data exfiltration capabilities were detected.
- [PROMPT_INJECTION]: No malicious instructions or bypass patterns (e.g., 'ignore previous instructions') are present in the skill's workflow or metadata.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text provided by the user (Category 8 surface). The risk is mitigated by the skill's analytical approach, which quantifies linguistic mechanics rather than executing the content's semantic meaning. Ingestion points: user-pasted samples in SKILL.md workflow. Capability inventory: Read/Write local social-context.md. Boundary markers: Not explicitly defined for samples. Sanitization: Implicitly managed by the quantitative nature of the analysis.
Audit Metadata