ludamo

Pass

Audited by Gen Agent Trust Hub on Jul 4, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill functions as a router that identifies user intent and interpolates user input into prompt templates for sub-skills. This creates a surface for indirect prompt injection where user-provided text is passed to subsequent logic layers.
  • Ingestion points: The skill captures user input in the '用户原话' section.
  • Boundary markers: The skill uses simple labels but lacks robust system-level delimiters or 'ignore nested instruction' directives.
  • Capability inventory: The instructions command the agent to read local context files ('读取') and switch to different behavioral protocols.
  • Sanitization: No specific input filtering or sanitization is mentioned before forwarding user text.
  • [DATA_EXPOSURE]: The skill instructs the agent to access local files via relative paths (e.g., '../ludamo-xuexi/SKILL.md') to load necessary context. This is a functional requirement of the routing logic rather than a data exfiltration attempt.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 4, 2026, 03:13 PM
Security Audit — agent-trust-hub — ludamo