From larkskill
Use this skill when operating Lark Approval via LarkSkill MCP: approval instance and task management.
How this skill is triggered — by the user, by Claude, or both
Slash command
/larkskill:lark-approvalThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
**CRITICAL — Before starting, MUST use the Read tool to read [`../lark-shared/SKILL.md`](../lark-shared/SKILL.md), which covers authentication and permission handling.**
CRITICAL — Before starting, MUST use the Read tool to read ../lark-shared/SKILL.md, which covers authentication and permission handling.
Use the LarkSkill MCP tool to call approval operations:
lark_api({ tool: 'approval', op: 'instances.get', args: { ... } })
lark_api_search({ query: 'approval instances', domain: 'approval' }) // discover available ops
Important: always use
lark_api_searchto discover the exact args shape for an operation before invoking it — do not guess field formats.
instances.get — Get details of a single approval instanceinstances.cancel — Cancel an approval instanceinstances.cc — CC an approval instanceinstances.initiated — Query the list of approval instances initiated by a usertasks.remind — Send a reminder to an approvertasks.approve — Approve an approval tasktasks.reject — Reject an approval tasktasks.transfer — Transfer an approval tasktasks.query — Query the task list for a usertasks.add_sign — Add a countersignature to an approval tasktasks.rollback — Roll back an approval task| Method | Required scope |
|---|---|
instances.get | approval:instance:read |
instances.cancel | approval:instance:write |
instances.cc | approval:instance:write |
instances.initiated | approval:instance:read |
tasks.remind | approval:instance:write |
tasks.approve | approval:task:write |
tasks.reject | approval:task:write |
tasks.transfer | approval:task:write |
tasks.query | approval:task:read |
tasks.add_sign | approval:task:write |
tasks.rollback | approval:task:write |
npx claudepluginhub kescyz/larkskill --plugin larkskillCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.