router

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill performs automated engine detection and task classification by reading untrusted project files such as project.godot, package.json, and Cargo.toml. This ingestion of external data creates a surface for indirect prompt injection if the files contain malicious instructions meant to be executed by the agent.
  • Ingestion points: The agent is instructed to scan project files and metadata to identify the engine and task discipline (SKILL.md, references/engine-detection.md).
  • Boundary markers: The skill logic lacks specific instructions or delimiters to isolate untrusted file content, making it possible for the agent to mistakenly follow instructions found within these files.
  • Capability inventory: The skill facilitates the loading of other specialized skills, which may grant the agent access to shell execution, network operations, or file system modifications.
  • Sanitization: There is no evidence of content validation or sanitization applied to the project files before they are processed by the routing algorithm.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 09:30 PM
Security Audit — agent-trust-hub — router