lark-workflow-onboarding
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Lark environment (such as user profiles, document titles, and chat names) and interpolates this content into subsequent tool commands and messages.
- Ingestion points: Results from
lark-cli contact +search-user,lark-cli wiki spaces nodes search, andlark-cli im +chats-searchin SKILL.md. - Boundary markers: The skill does not use specific delimiters or instructions to prevent the agent from following potential commands embedded within the retrieved data.
- Capability inventory: The skill has write capabilities across multiple domains, including sending IM messages, creating tasks, and scheduling calendar events via
lark-cli. - Sanitization: No sanitization or validation of the data retrieved from external API calls is performed before interpolation.
- [EXTERNAL_DOWNLOADS]: The skill references external configuration and security documentation hosted on the official GitHub repository for the Lark Suite CLI.
- References: The skill points to documentation at
https://github.com/larksuite/cli/for authentication and safety rules.
Audit Metadata