lark-event
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes real-time event data from Lark/Feishu, which includes untrusted external inputs like IM messages, meeting topics, and task descriptions.
- Ingestion points: Data enters the agent context through the
lark-cli event consumecommand across multiple domains including IM, Approval, Task, VC, and Minutes. - Boundary markers: The skill documentation does not provide instructions for the agent to use explicit delimiters or escaping mechanisms when processing event content, which could lead to instructions in the events being interpreted as agent commands.
- Capability inventory: The skill allows the agent to execute the
lark-clibinary, perform data transformations usingjqexpressions, and write event data to the local file system. - Sanitization: The skill implements a security control for the
--output-dirflag, restricting file writes to relative paths to mitigate directory traversal risks.
Audit Metadata