The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill retrieves and summarizes content from a knowledge base. This creates a surface where instructions embedded within those files could influence the agent's reasoning or output. \n- Ingestion points: Data is pulled from local files via the Read tool. \n- Boundary markers: No delimiters or explicit 'ignore instructions' warnings are defined to separate knowledge base content from system instructions. \n- Capability inventory: The skill is restricted to the Read tool; it lacks network access or file-writing capabilities, which significantly limits the risk. \n- Sanitization: No sanitization of retrieved data is performed. \n- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials or unauthorized data transmission patterns were found. \n- [Remote Code Execution] (SAFE): The skill does not download or execute remote scripts or external packages.

knowledge-query