lark-workflow-multi-agent-dev
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
lark-clito interact with Feishu APIs for table creation, record management, and messaging. These are legitimate uses of the specified binary dependency in the frontmatter. - [COMMAND_EXECUTION]: The skill orchestrates multiple sub-agents using a native
Agent()tool provided by the underlying AI platform (Claude Code). It describes patterns for 'Competition', 'Division of Labor', and 'Pipeline' modes using these sub-agents. - [COMMAND_EXECUTION]: Integration with standard developer tools like
gitandgh(GitHub CLI) for repository operations is documented as part of the intended workflow. - [CREDENTIALS_SAFE]: The skill references an external
lark-sharedskill for authentication and instructs the user to login vialark-cli auth login, following safe secret management practices without hardcoding sensitive data.
Audit Metadata