dingtalk-doc
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill operates through the
dwsCLI to perform document management tasks. It includes strict rules against guessing command flags or parameters and requires the use of structured output (--format json) for reliable parsing. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to read, analyze, and modify external document content. This creates a surface for indirect prompt injection where malicious instructions embedded in a document could attempt to influence the agent. The skill mitigates this by enforcing structured data processing (JSONML) and requiring explicit verification of written content.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill contains explicit safety instructions (Runtime Contract) to prevent the leakage of sensitive information. It forbids the agent from outputting or recording authentication tokens, refresh tokens, appSecrets, or webhook tokens. It also restricts file operations to the current working directory using relative paths (
@相对路径) to prevent directory traversal attacks.
Audit Metadata