collaborating-with-kimi
Skill1 starsChanged 8 months agoMITView on GitHub
[由 collaborating-hub 路由] Kimi CLI 后端实现。直接使用请通过 /collab 或 /kimi 触发 collaborating-hub。
In this repository10 other files it gives its agents.
- skills/agent-memory/CLAUDE.mdCLAUDE.md
- agent-memorySkill
- auto-iterateSkill
- collaborating-with-codexSkill
- collaborating-with-geminiSkill
- explorationSkill
- reflex-auditSkill
- reflex-fuzzSkill
- skill-catalogSkill
- skill-snapshotSkill
The file
Files
Quick Start
~/.claude/skills/collaborating-with-kimi/scripts/kimi-agent.sh "你的任务" --cd "/path/to/project"输出: JSON 格式,包含 success, session_id, response, 可选 error。
Parameters
用法: kimi-agent.sh "任务描述" --cd /path/to/project [选项]
必需参数:
--cd PATH 项目工作目录
可选参数:
--agent, -a NAME 注入 agent 提示词 (如 security-reviewer, planner)
--session, -s ID 续接会话 (使用之前返回的 session_id)
--context, -c FILE 自定义上下文文件
--model, -m MODEL 指定模型 (默认使用 kimi 配置)
--timeout, -t SECS 超时时间 (默认 120 秒)
--yolo 自动批准所有操作
--thinking 启用思考模式
--no-thinking 禁用思考模式
--output-format 输出格式 (json|text)
--save 保存会话到文件 (默认开启)
--no-save 不保存会话Multi-turn Sessions
始终捕获 session_id 用于后续对话:
# 初始任务
~/.claude/skills/collaborating-with-kimi/scripts/kimi-agent.sh "分析 login.py 的认证逻辑" --cd "/project"
# 使用 session_id 继续
~/.claude/skills/collaborating-with-kimi/scripts/kimi-agent.sh "为刚才分析的代码写单元测试" --cd "/project" --session "kimi_20260128_123456_12345"Common Patterns
代码审查 (只读模式):
kimi-agent.sh "审查这段代码的安全性" --cd "/project" --agent security-reviewer规划任务:
kimi-agent.sh "规划实现用户认证功能" --cd "/project" --agent planner自动执行 (YOLO 模式):
kimi-agent.sh "重构这个函数" --cd "/project" --yolo启用深度思考:
kimi-agent.sh "复杂架构分析" --cd "/project" --thinkingSession Management
会话自动保存到 ~/.claude/skills/collaborating-with-kimi/sessions/
使用会话管理工具:
# 列出所有会话
~/.claude/skills/collaborating-with-kimi/scripts/kimi-sessions.sh list
# 查看会话详情
~/.claude/skills/collaborating-with-kimi/scripts/kimi-sessions.sh show <session_id>
# 清理 7 天前的会话
~/.claude/skills/collaborating-with-kimi/scripts/kimi-sessions.sh clean 7Context Injection
自动加载上下文优先级:
--context FILE指定的文件$PROJECT_DIR/CLAUDE.md$PROJECT_DIR/AGENTS.md$PROJECT_DIR/KIMI.md
Agent Injection
可用的 agents(位于 ~/.claude/agents/):
planner- 实现规划architect- 系统设计security-reviewer- 安全审查code-reviewer- 代码审查tdd-guide- 测试驱动开发- 更多...
Output Format
成功响应:
{
"success": true,
"session_id": "kimi_20260128_123456_12345",
"agent": "security-reviewer",
"context_source": "CLAUDE.md",
"response": "Kimi 的响应内容..."
}失败响应:
{
"success": false,
"session_id": "kimi_20260128_123456_12345",
"error": "错误信息",
"exit_code": 1
}Troubleshooting
Kimi 执行超时:
- 增加
--timeout值 - 检查网络连接
- 确认 Kimi CLI 认证有效:
kimi info
无响应输出:
- 确保 Kimi CLI 已正确安装和认证
- 尝试在终端直接运行
kimi -p "test"测试
登录问题:
- 运行
kimi login重新认证
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

