lark-event
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on the
lark-clibinary to perform all actions, including listing event types, retrieving schemas, and establishing long-running connections to consume event streams as NDJSON. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, untrusted sources within the Lark environment, such as IM message content, document comments, and task updates. Malicious content within these events could potentially influence the behavior of the AI agent if not properly handled.
- Ingestion points: Data enters the agent's context through
lark-cli event consumeoutput (stdout), specifically within fields like.contentinim.message.receive_v1or comment strings in Drive events. - Boundary markers: The instructions do not explicitly define boundary markers or delimiters to isolate the untrusted event data from the agent's operational instructions.
- Capability inventory: The agent possesses capabilities provided by
lark-cliand potentially other tools in thelark-sharedcontext, including the ability to send messages, read documents, and manage platform resources. - Sanitization: There is no evidence of sanitization or filtering of external content described in the skill instructions, although it does note that
lark-clipre-renders some content (like resolving @mentions) for readability.
Audit Metadata