lark-workflow-daily-briefing
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill aggregates untrusted data from multiple sources, which could contain malicious instructions designed to influence the AI's behavior.
- Ingestion points: The workflow retrieves data from mail, IM messages, tasks, and approvals via
lark-cliin Steps 3 through 6. - Boundary markers: Step 7 (AI summarization) does not define any delimiters or safety instructions to separate the untrusted data from the summary instructions.
- Capability inventory: The skill has the capability to write data back to the ecosystem via
lark-cli docs +create. - Sanitization: No sanitization, escaping, or filtering of the external content is documented before it is passed to the AI.
- [COMMAND_EXECUTION]: The skill relies on local command-line tools for its primary functionality.
- Invokes
lark-clifor all data retrieval and document creation tasks across business domains. - Uses the system
datecommand for ISO 8601 timestamp calculations. - [EXTERNAL_DOWNLOADS]: The skill references external resources for configuration and logic.
- Fetches shared skill definitions and security rules from the larksuite organization's official GitHub repository.
Audit Metadata