deepseek-helper

Pass

Audited by Gen Agent Trust Hub on May 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists of informational content intended to guide users in using the DeepSeek API. No executable scripts, command execution, or sensitive data access patterns were found.
  • [SAFE]: All external links point to the official DeepSeek platform and API documentation, posing no security risk.
  • [SAFE]: Indirect Prompt Injection Surface: The skill processes user queries as ingestion points. No explicit boundary markers or sanitization are defined in the instructions. However, the capability inventory across the skill shows no dangerous operations such as file-writing, network tool usage, or shell execution, neutralizing the impact of potential injections.
Audit Metadata
Risk Level
SAFE
Analyzed
May 6, 2026, 04:19 PM
Security Audit — agent-trust-hub — deepseek-helper