dingtalk-document
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to generate temporary shell scripts in
/tmp/for complex API operations. This is a standard functional pattern for handling multi-line logic and variable interpolation in this environment. - [DYNAMIC_EXECUTION]: Shell scripts are dynamically constructed by the agent based on API documentation and user requirements, then executed to interact with DingTalk services.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it retrieves and processes content from external DingTalk API endpoints.
- Ingestion points: Data returned from DingTalk API calls, such as document bodies and workspace metadata.
- Boundary markers: The skill does not define specific delimiters for external data, relying on the agent's internal processing of JSON structures.
- Capability inventory: Includes shell command execution, file system access (via helper script), and network operations.
- Sanitization: No explicit sanitization logic is provided in the instructions for the external data being processed.
Audit Metadata