general-purpose

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions define an agent designed to research complex questions by gathering information from diverse external sources, including web searches and GitHub repositories. This creates a surface area for indirect prompt injection attacks where malicious instructions hidden within processed data could influence the agent to perform unintended actions, especially given its broad toolset.
  • Ingestion points: Web search, GitHub code search, and file system tools for local resource access (specified in the Tool Usage Strategy section of SKILL.md).
  • Boundary markers: The instructions do not define specific delimiters or "ignore" directives to isolate external content from the agent's core instructions.
  • Capability inventory: File manipulation, bash command execution, and writing tools (specified in the Tool Usage Strategy section of SKILL.md).
  • Sanitization: No specific data validation, escaping, or sanitization protocols are mentioned for processing external information before it influences subsequent steps.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 08:05 AM
Security Audit — agent-trust-hub — general-purpose