skills/vm0-ai/vm0-skills/lark/Gen Agent Trust Hub

lark

Pass

Audited by Gen Agent Trust Hub on Apr 6, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses curl and jq to interact with the Lark/Feishu API. This includes operations for sending messages, managing groups, and querying directory structure. All network requests are directed to official open.feishu.cn or open.larkoffice.com endpoints.
  • [DATA_EXPOSURE]: The skill includes a helper function that caches the tenant_access_token in /tmp/lark_token.json. While this is a functional implementation for token management, /tmp is a shared directory on many systems.
  • [INDIRECT_PROMPT_INJECTION]: The skill has the capability to ingest untrusted data from external sources, specifically when fetching chat history via the im/v1/messages endpoint or searching for users. This represents an attack surface where malicious content from external users could influence agent behavior, though the skill provides guidance on using jq to escape content as a mitigation measure.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 6, 2026, 01:43 AM