design-lark-chart
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by processing external data from natural language demands and Feishu document URLs. * Ingestion points: Natural language requests and content retrieved from Feishu docx or wiki URLs are used to extract semantic data for planning. * Boundary markers: There are no explicit instructions or delimiters used to ensure the agent ignores instructions embedded within user-provided or document-retrieved text. * Capability inventory: The skill can modify whiteboard content via 'lark-cli whiteboard +update' and execute geometric checks. * Sanitization: Risk is mitigated through a mandatory Visual Quality Assurance loop requiring two independent reviewers to score output based on the plan.
- [EXTERNAL_DOWNLOADS]: The skill downloads and executes the '@larksuite/whiteboard-cli' package from the npm registry via npx to perform diagram validation and rendering.
- [COMMAND_EXECUTION]: Several scripts execute command-line utilities including 'lark-cli' and 'whiteboard-cli' to interact with the Feishu platform and validate generated whiteboard data.
Audit Metadata