lark-crm

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill interfaces with the Feishu (Lark) Open Platform by executing the lark-cli binary. It performs CRM management tasks through commands like +record-list, +record-upsert, and +data-query to handle database records.
  • [EXTERNAL_DOWNLOADS]: The README file provides a link to a WeChat (mp.weixin.qq.com) article containing installation instructions for the lark-cli tool. WeChat is an established and well-known service provider.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, specifically user-provided URLs and records retrieved from the CRM system.
  • Ingestion points: The skill extracts resource identifiers (base_token) from URLs provided by the user and reads data from CRM records via API calls.
  • Boundary markers: The instructions do not implement specific delimiters or warnings to isolate potentially malicious instructions embedded within the ingested data.
  • Capability inventory: The skill possesses the ability to execute system commands through the CLI and write or update data within the CRM platform.
  • Sanitization: The skill relies on natural language extraction rather than programmatic validation to handle inputs interpolated into the CLI commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 04:15 PM
Security Audit — agent-trust-hub — lark-crm