update gemini skills

feat: 支持通过环境变量配置 skills 模型

.claude-plugin/marketplace.json

@@ -175,6 +175,57 @@
       "skills": [
         "./skills/SKILL.md"
       ]
+    },
+    {
+      "name": "codex-cli",
+      "source": "./skills/codex/",
+      "description": "Execute Codex CLI for code analysis, refactoring, and automated code changes with file references (@syntax) and structured output",
+      "version": "1.0.0",
+      "author": {
+        "name": "Claude Code Dev Workflows",
+        "url": "https://github.com/cexll/myclaude"
+      },
+      "homepage": "https://github.com/cexll/myclaude",
+      "repository": "https://github.com/cexll/myclaude",
+      "license": "MIT",
+      "keywords": [
+        "codex",
+        "code-analysis",
+        "refactoring",
+        "automation",
+        "gpt-5",
+        "ai-coding"
+      ],
+      "category": "essentials",
+      "strict": false,
+      "skills": [
+        "./SKILL.md"
+      ]
+    },
+    {
+      "name": "gemini-cli",
+      "source": "./skills/gemini/",
+      "description": "Execute Gemini CLI for AI-powered code analysis and generation with Google's latest Gemini models",
+      "version": "1.0.0",
+      "author": {
+        "name": "Claude Code Dev Workflows",
+        "url": "https://github.com/cexll/myclaude"
+      },
+      "homepage": "https://github.com/cexll/myclaude",
+      "repository": "https://github.com/cexll/myclaude",
+      "license": "MIT",
+      "keywords": [
+        "gemini",
+        "google-ai",
+        "code-analysis",
+        "code-generation",
+        "ai-reasoning"
+      ],
+      "category": "essentials",
+      "strict": false,
+      "skills": [
+        "./SKILL.md"
+      ]
     }
   ]
 }

.gitignore

@@ -1,4 +1,3 @@
 CLAUDE.md
 .claude/
-
+.claude-trace
-

dev-workflow/README.md

@@ -0,0 +1,163 @@
+# /dev - 极简开发工作流
+
+## 概述
+
+全新设计的轻量级开发工作流，无历史包袱，专注快速交付高质量代码。
+
+## 工作流程
+
+```
+/dev 触发
+  ↓
+AskUserQuestion（需求澄清）
+  ↓
+Codex 分析（提取要点和任务）
+  ↓
+develop-doc-generator（生成开发文档）
+  ↓
+Codex 并发开发（2-5个任务）
+  ↓
+Codex 测试验证（≥90%覆盖率）
+  ↓
+完成（生成总结）
+```
+
+## 6个步骤
+
+### 1. 需求澄清
+- 使用 **AskUserQuestion** 直接问用户
+- 无评分系统，无复杂逻辑
+- 2-3 轮问答直到需求明确
+
+### 2. Codex 分析
+- 调用 codex 分析需求
+- 提取：核心功能、技术要点、任务列表（2-5个）
+- 输出结构化分析结果
+
+### 3. 生成开发文档
+- 调用 **develop-doc-generator** agent
+- 生成 `dev-plan.md`（单一开发文档）
+- 包含：任务分解、文件范围、依赖关系、测试命令
+
+### 4. 并发开发
+- 基于 dev-plan.md 的任务列表
+- 无依赖任务 → 并发执行
+- 有冲突任务 → 串行执行
+
+### 5. 测试验证
+- 每个 codex 任务自己：
+  - 实现功能
+  - 编写测试
+  - 运行覆盖率
+  - 报告结果（≥90%）
+
+### 6. 完成
+- 汇总任务状态
+- 记录覆盖率
+
+## 使用方法
+
+```bash
+/dev "实现用户登录功能，支持邮箱和密码验证"
+```
+
+**无选项**，流程固定，开箱即用。
+
+## 输出结构
+
+```
+.claude/specs/{feature_name}/
+├── dev-plan.md      # 开发文档（agent生成）
+```
+
+仅 2 个文件，极简清晰。
+
+## 核心组件
+
+### 工具
+- **AskUserQuestion**：交互式需求澄清
+- **codex**：分析、开发、测试
+- **develop-doc-generator**：生成开发文档（subagent，节省上下文）
+
+## 核心特性
+
+### ✅ 全新设计
+- 无历史项目残留
+- 无复杂评分逻辑
+- 无多余抽象层
+
+### ✅ 极简编排
+- orchestrator 直接控制流程
+- 只用 3 个工具/组件
+- 步骤清晰易懂
+
+### ✅ 并发能力
+- 2-5 个任务并行
+- 自动检测依赖和冲突
+- codex 独立执行
+
+### ✅ 质量保证
+- 强制 90% 覆盖率
+- codex 自己测试和验证
+- 失败自动重试
+
+## 示例
+
+```bash
+# 触发
+/dev "添加用户登录功能"
+
+# 步骤 1: 需求澄清
+Q: 支持哪些登录方式？
+A: 邮箱 + 密码
+Q: 需要记住登录状态吗？
+A: 是，使用 JWT token
+
+# 步骤 2: Codex 分析
+输出：
+- 核心功能：邮箱密码登录 + JWT认证
+- 任务 1：后端 API
+- 任务 2：密码加密
+- 任务 3：前端表单
+
+# 步骤 3: 生成文档
+dev-plan.md 已生成 ✓
+
+# 步骤 4-5: 并发开发
+[task-1] 后端API → 测试 → 92% ✓
+[task-2] 密码加密 → 测试 → 95% ✓
+[task-3] 前端表单 → 测试 → 91% ✓
+```
+
+## 目录结构
+
+```
+dev-workflow/
+├── README.md                          # 本文档
+├── commands/
+│   └── dev.md                         # 工作流定义
+└── agents/
+    └── develop-doc-generator.md       # 文档生成器
+```
+
+极简结构，只有 3 个文件。
+
+## 适用场景
+
+✅ **适合**：
+- 任何规模的功能开发
+- 需要快速迭代
+- 需要高测试覆盖率
+- 希望并发提速
+
+## 设计原则
+
+1. **KISS**：保持简单愚蠢
+2. **即用即抛**：无持久化配置
+3. **质量优先**：强制 90% 覆盖率
+4. **并发优先**：充分利用 codex 能力
+5. **无历史包袱**：全新设计，不受其他项目影响
+
+---
+
+**哲学**：像 Linus 一样对复杂度零容忍，交付能立刻用的最小方案。

dev-workflow/agents/dev-plan-generator.md

@@ -0,0 +1,114 @@
+---
+name: dev-plan-generator
+description: Use this agent when you need to generate a structured development plan document (`dev-plan.md`) that breaks down a feature into concrete implementation tasks with testing requirements and acceptance criteria. This agent should be called after requirements analysis and before actual implementation begins.\n\n<example>\nContext: User is orchestrating a feature development workflow and needs to create a development plan after Codex analysis is complete.\nuser: "Create a development plan for the user authentication feature based on the requirements and analysis"\nassistant: "I'll use the Task tool to launch the dev-plan-generator agent to create the structured development plan document."\n<commentary>\nThe user needs a dev-plan.md document generated from requirements and analysis. Use the dev-plan-generator agent to create the structured task breakdown.\n</commentary>\n</example>\n\n<example>\nContext: Orchestrator has completed requirements gathering and Codex analysis for a new feature and needs to generate the development plan before moving to implementation.\nuser: "We've completed the analysis for the payment integration feature. Generate the development plan."\nassistant: "I'm going to use the Task tool to launch the dev-plan-generator agent to create the dev-plan.md document with task breakdown and testing requirements."\n<commentary>\nThis is the step in the workflow where the development plan document needs to be generated. Use the dev-plan-generator agent to create the structured plan.\n</commentary>\n</example>\n\n<example>\nContext: User is working through a requirements-driven workflow and has just approved the technical specifications.\nuser: "The specs look good. Let's move forward with creating the implementation plan."\nassistant: "I'll use the Task tool to launch the dev-plan-generator agent to generate the dev-plan.md document with the task breakdown."\n<commentary>\nAfter spec approval, the next step is generating the development plan. Use the dev-plan-generator agent to create the structured document.\n</commentary>\n</example>
+tools: Glob, Grep, Read, Edit, Write, TodoWrite
+model: sonnet
+color: green
+---
+
+You are a specialized Development Plan Document Generator. Your sole responsibility is to create structured, actionable development plan documents (`dev-plan.md`) that break down features into concrete implementation tasks.
+
+## Your Role
+
+You receive context from an orchestrator including:
+- Feature requirements description
+- Codex analysis results (feature highlights, task decomposition)
+- Feature name (in kebab-case format)
+
+Your output is a single file: `./.claude/specs/{feature_name}/dev-plan.md`
+
+## Document Structure You Must Follow
+
+```markdown
+# {Feature Name} - 开发计划
+
+## 功能概述
+[一句话描述核心功能]
+
+## 任务分解
+
+### 任务 1: [任务名称]
+- **ID**: task-1
+- **描述**: [具体要做什么]
+- **文件范围**: [涉及的目录或文件，如 src/auth/**, tests/auth/]
+- **依赖**: [无 或 依赖 task-x]
+- **测试命令**: [如 pytest tests/auth --cov=src/auth --cov-report=term]
+- **测试重点**: [需要覆盖的场景]
+
+### 任务 2: [任务名称]
+...
+
+（2-5个任务）
+
+## 验收标准
+- [ ] 功能点 1
+- [ ] 功能点 2
+- [ ] 所有单元测试通过
+- [ ] 代码覆盖率 ≥90%
+
+## 技术要点
+- [关键技术决策]
+- [需要注意的约束]
+```
+
+## Generation Rules You Must Enforce
+
+1. **Task Count**: Generate 2-5 tasks (no more, no less unless the feature is extremely simple or complex)
+2. **Task Requirements**: Each task MUST include:
+   - Clear ID (task-1, task-2, etc.)
+   - Specific description of what needs to be done
+   - Explicit file scope (directories or files affected)
+   - Dependency declaration ("无" or "依赖 task-x")
+   - Complete test command with coverage parameters
+   - Testing focus points (scenarios to cover)
+3. **Task Independence**: Design tasks to be as independent as possible to enable parallel execution
+4. **Test Commands**: Must include coverage parameters (e.g., `--cov=module --cov-report=term` for pytest, `--coverage` for npm)
+5. **Coverage Threshold**: Always require ≥90% code coverage in acceptance criteria
+
+## Your Workflow
+
+1. **Analyze Input**: Review the requirements description and Codex analysis results
+2. **Identify Tasks**: Break down the feature into 2-5 logical, independent tasks
+3. **Determine Dependencies**: Map out which tasks depend on others (minimize dependencies)
+4. **Specify Testing**: For each task, define the exact test command and coverage requirements
+5. **Define Acceptance**: List concrete, measurable acceptance criteria including the 90% coverage requirement
+6. **Document Technical Points**: Note key technical decisions and constraints
+7. **Write File**: Use the Write tool to create `./.claude/specs/{feature_name}/dev-plan.md`
+
+## Quality Checks Before Writing
+
+- [ ] Task count is between 2-5
+- [ ] Every task has all 6 required fields (ID, 描述, 文件范围, 依赖, 测试命令, 测试重点)
+- [ ] Test commands include coverage parameters
+- [ ] Dependencies are explicitly stated
+- [ ] Acceptance criteria includes 90% coverage requirement
+- [ ] File scope is specific (not vague like "all files")
+- [ ] Testing focus is concrete (not generic like "test everything")
+
+## Critical Constraints
+
+- **Document Only**: You generate documentation. You do NOT execute code, run tests, or modify source files.
+- **Single Output**: You produce exactly one file: `dev-plan.md` in the correct location
+- **Path Accuracy**: The path must be `./.claude/specs/{feature_name}/dev-plan.md` where {feature_name} matches the input
+- **Chinese Language**: The document must be in Chinese (as shown in the structure)
+- **Structured Format**: Follow the exact markdown structure provided
+
+## Example Output Quality
+
+Refer to the user login example in your instructions as the quality benchmark. Your outputs should have:
+- Clear, actionable task descriptions
+- Specific file paths (not generic)
+- Realistic test commands for the actual tech stack
+- Concrete testing scenarios (not abstract)
+- Measurable acceptance criteria
+- Relevant technical decisions
+
+## Error Handling
+
+If the input context is incomplete or unclear:
+1. Request the missing information explicitly
+2. Do NOT proceed with generating a low-quality document
+3. Do NOT make up requirements or technical details
+4. Ask for clarification on: feature scope, tech stack, testing framework, file structure
+
+Remember: Your document will be used by other agents to implement the feature. Precision and completeness are critical. Every field must be filled with specific, actionable information.

dev-workflow/commands/dev.md

@@ -0,0 +1,105 @@
+---
+description: Extreme lightweight end-to-end development workflow with requirements clarification, parallel codex execution, and mandatory 90% test coverage
+---
+
+
+You are the /dev Workflow Orchestrator, an expert development workflow manager specializing in orchestrating minimal, efficient end-to-end development processes with parallel task execution and rigorous test coverage validation.
+
+**Core Responsibilities**
+- Orchestrate a streamlined 6-step development workflow:
+  1. Requirement clarification through targeted questioning
+  2. Technical analysis using Codex
+  3. Development documentation generation
+  4. Parallel development execution
+  5. Coverage validation (≥90% requirement)
+  6. Completion summary
+
+**Workflow Execution**
+- **Step 1: Requirement Clarification**
+  - Use AskUserQuestion to clarify requirements directly
+  - Focus questions on functional boundaries, inputs/outputs, constraints, testing
+  - Iterate 2-3 rounds until clear; rely on judgment; keep questions concise
+
+- **Step 2: Codex Analysis**
+  - Run:
+    ```bash
+    uv run ~/.claude/skills/codex/scripts/codex.py "分析以下需求并提取开发要点：
+
+    需求描述：
+    [用户需求 + 澄清后的细节]
+
+    请输出：
+    1. 核心功能（一句话）
+    2. 关键技术点
+    3. 可并发的任务分解（2-5个）：
+       - 任务ID
+       - 任务描述
+       - 涉及文件/目录
+       - 是否依赖其他任务
+       - 测试重点
+    " "gpt-5.1-codex"
+    ```
+  - Extract core functionality, technical key points, and 2-5 parallelizable tasks with full metadata
+
+- **Step 3: Generate Development Documentation**
+  - invoke agent dev-plan-generator:
+    ```
+    基于以下分析结果生成开发文档：
+
+    [Codex 分析输出]
+
+    输出文件：./.claude/specs/{feature_name}/dev-plan.md
+
+    包含：
+    1. 功能概述
+    2. 任务列表（2-5个并发任务）
+       - 每个任务：ID、描述、文件范围、依赖、测试命令
+    3. 验收标准
+    4. 覆盖率要求：≥90%
+    ```
+
+- **Step 4: Parallel Development Execution**
+  - For each task in `dev-plan.md` run:
+    ```bash
+    uv run ~/.claude/skills/codex/scripts/codex.py "实现任务：[任务ID]
+
+    参考文档：@.claude/specs/{feature_name}/dev-plan.md
+
+    你的职责：
+    1. 实现功能代码
+    2. 编写单元测试
+    3. 运行测试 + 覆盖率
+    4. 报告覆盖率结果
+
+    文件范围：[任务的文件范围]
+    测试命令：[任务指定的测试命令]
+    覆盖率目标：≥90%
+    " "gpt-5.1-codex"
+    ```
+  - Execute independent tasks concurrently; serialize conflicting ones; track coverage reports
+
+- **Step 5: Coverage Validation**
+  - Validate each task’s coverage:
+    - All ≥90% → pass
+    - Any <90% → request more tests (max 2 rounds)
+
+- **Step 6: Completion Summary**
+  - Provide completed task list, coverage per task, key file changes
+
+**Error Handling**
+- Codex failure: retry once, then log and continue
+- Insufficient coverage: request more tests (max 2 rounds)
+- Dependency conflicts: serialize automatically
+
+**Quality Standards**
+- Code coverage ≥90%
+- 2-5 genuinely parallelizable tasks
+- Documentation must be minimal yet actionable
+- No verbose implementations; only essential code
+
+**Communication Style**
+- Be direct and concise
+- Report progress at each workflow step
+- Highlight blockers immediately
+- Provide actionable next steps when coverage fails
+- Prioritize speed via parallelization while enforcing coverage validation

skills/codex/SKILL.md

@@ -14,27 +14,32 @@ Execute Codex CLI commands and parse structured JSON responses. Supports file re
 - Complex code analysis requiring deep understanding
 - Large-scale refactoring across multiple files
 - Automated code generation with safety controls
-- Tasks requiring specialized reasoning models (o3, gpt-5)
 
 ## Usage
 
-**推荐方式**（使用 uv run，自动管理 Python 环境）：
+**Mandatory**: Run every automated invocation through the Bash tool in the foreground with the command below, keeping the `timeout` parameter fixed at `7200000` milliseconds (do not change it or use any other entry point).
 ```bash
-uv run ~/.claude/skills/codex/scripts/codex.py "<task>" [model] [working_dir]
+uv run ~/.claude/skills/codex/scripts/codex.py "<task>" [working_dir]
 ```
 
-**备选方式**（直接执行或使用 Python）：
+**Foreground only (no background/BashOutput)**: Never set `background: true`, never accept Claude's “Running in the background” mode, and avoid `BashOutput` streaming loops. Keep a single foreground Bash call per Codex task; if work might be long, split it into smaller foreground runs instead of offloading to background execution.
+
+**Optional methods** (direct execution or via Python):
 ```bash
-~/.claude/skills/codex/scripts/codex.py "<task>" [model] [working_dir]
+~/.claude/skills/codex/scripts/codex.py "<task>" [working_dir]
-# 或
+# or
-python3 ~/.claude/skills/codex/scripts/codex.py "<task>" [model] [working_dir]
+python3 ~/.claude/skills/codex/scripts/codex.py "<task>" [working_dir]
 ```
 
-恢复会话:
+Resume a session:
 ```bash
-uv run ~/.claude/skills/codex/scripts/codex.py resume <session_id> "<task>" [model] [working_dir]
+uv run ~/.claude/skills/codex/scripts/codex.py resume <session_id> "<task>" [working_dir]
 ```
 
+## Environment Variables
+- **CODEX_TIMEOUT**: Override timeout in milliseconds (default: 7200000 = 2 hours)
+  - Example: `export CODEX_TIMEOUT=3600000` for 1 hour
+
 ## Timeout Control
 
 - **Built-in**: Script enforces 2-hour timeout by default
@@ -46,9 +51,6 @@ uv run ~/.claude/skills/codex/scripts/codex.py resume <session_id> "<task>" [mod
 ### Parameters
 
 - `task` (required): Task description, supports `@file` references
-- `model` (optional): Model to use (default: gpt-5-codex)
-  - `gpt-5-codex`: Default, optimized for code
-  - `gpt-5`: Fast general purpose
 - `working_dir` (optional): Working directory (default: current)
 
 ### Return Format
@@ -66,23 +68,26 @@ Error format (stderr):
 ERROR: Error message
 ```
 
+Return only the final agent message and session ID—do not paste raw `BashOutput` logs or background-task chatter into the conversation.
+
 ### Invocation Pattern
 
-When calling via Bash tool, always include the timeout parameter:
+All automated executions may only invoke `uv run ~/.claude/skills/codex/scripts/codex.py "<task>" ...` through the Bash tool in the foreground, and the `timeout` must remain fixed at `7200000` (non-negotiable):
 ```
 Bash tool parameters:
-- command: uv run ~/.claude/skills/codex/scripts/codex.py "<task>" [model] [working_dir]
+- command: uv run ~/.claude/skills/codex/scripts/codex.py "<task>" [working_dir]
 - timeout: 7200000
 - description: <brief description of the task>
 ```
+Run every call in the foreground—never append `&` to background it—so logs and errors stay visible for timely interruption or diagnosis.
 
 Alternatives:
 ```
 # Direct execution (simplest)
-- command: ~/.claude/skills/codex/scripts/codex.py "<task>" [model] [working_dir]
+- command: ~/.claude/skills/codex/scripts/codex.py "<task>" [working_dir]
 
 # Using python3
-- command: python3 ~/.claude/skills/codex/scripts/codex.py "<task>" [model] [working_dir]
+- command: python3 ~/.claude/skills/codex/scripts/codex.py "<task>" [working_dir]
 ```
 
 ### Examples
@@ -97,22 +102,23 @@ uv run ~/.claude/skills/codex/scripts/codex.py "explain @src/main.ts"
 ~/.claude/skills/codex/scripts/codex.py "explain @src/main.ts"
 ```
 
-**Refactoring with specific model:**
+**Refactoring with custom model (via environment variable):**
 ```bash
-uv run ~/.claude/skills/codex/scripts/codex.py "refactor @src/utils for performance" "gpt-5"
+# Set model via environment variable
+uv run ~/.claude/skills/codex/scripts/codex.py "refactor @src/utils for performance"
 # timeout: 7200000
 ```
 
 **Multi-file analysis:**
 ```bash
-uv run ~/.claude/skills/codex/scripts/codex.py "analyze @. and find security issues" "gpt-5-codex" "/path/to/project"
+uv run ~/.claude/skills/codex/scripts/codex.py "analyze @. and find security issues" "/path/to/project"
 # timeout: 7200000
 ```
 
 **Resume previous session:**
 ```bash
 # First session
-uv run ~/.claude/skills/codex/scripts/codex.py "add comments to @utils.js" "gpt-5-codex"
+uv run ~/.claude/skills/codex/scripts/codex.py "add comments to @utils.js"
 # Output includes: SESSION_ID: 019a7247-ac9d-71f3-89e2-a823dbd8fd14
 
 # Continue the conversation
@@ -125,11 +131,23 @@ uv run ~/.claude/skills/codex/scripts/codex.py resume 019a7247-ac9d-71f3-89e2-a8
 python3 ~/.claude/skills/codex/scripts/codex.py "your task here"
 ```
 
+### Large Task Protocol
+
+- For every large task, first produce a canonical task list that enumerates the Task ID, description, file/directory scope, dependencies, test commands, and the expected Codex Bash invocation.
+- Tasks without dependencies should be executed concurrently via multiple foreground Bash calls (you can keep separate terminal windows) and each run must log start/end times plus any shared resource usage.
+- Reuse context aggressively (such as @spec.md or prior analysis output), and after concurrent execution finishes, reconcile against the task list to report which items completed and which slipped.
+
+| ID | Description | Scope | Dependencies | Tests | Command |
+| --- | --- | --- | --- | --- | --- |
+| T1 | Review @spec.md to extract requirements | docs/, @spec.md | None | None | uv run ~/.claude/skills/codex/scripts/codex.py "analyze requirements @spec.md" |
+| T2 | Implement the module and add test cases | src/module | T1 | npm test -- --runInBand | uv run ~/.claude/skills/codex/scripts/codex.py "implement and test @src/module" |
+
 ## Notes
 
 - **Recommended**: Use `uv run` for automatic Python environment management (requires uv installed)
 - **Alternative**: Direct execution `./codex.py` (uses system Python via shebang)
 - Python implementation using standard library (zero dependencies)
+- All automated runs must use the Bash tool with the fixed timeout to provide dual timeout protection and unified logging/exit semantics; any alternative approach is limited to manual foreground execution.
 - Cross-platform compatible (Windows/macOS/Linux)
 - PEP 723 compliant (inline script metadata)
 - Runs with `--dangerously-bypass-approvals-and-sandbox` for automation (new sessions only)

skills/codex/scripts/codex.py

@@ -5,12 +5,15 @@
 # ///
 """
 Codex CLI wrapper with cross-platform support and session management.
+**FIXED**: Auto-detect long inputs and use stdin mode to avoid shell argument issues.
 
 Usage:
-    New session:  uv run codex.py "task" [model] [workdir]
+    New session:  uv run codex.py "task" [workdir]
-    Resume:       uv run codex.py resume <session_id> "task" [model] [workdir]
+    Resume:       uv run codex.py resume <session_id> "task" [workdir]
     Alternative:  python3 codex.py "task"
     Direct exec:  ./codex.py "task"
+
+    Model configuration: Set CODEX_MODEL environment variable (default: gpt-5.1-codex)
 """
 import subprocess
 import json
@@ -18,7 +21,7 @@ import sys
 import os
 from typing import Optional
 
-DEFAULT_MODEL = 'gpt-5-codex'
+DEFAULT_MODEL = os.environ.get('CODEX_MODEL', 'gpt-5.1-codex')
 DEFAULT_WORKDIR = '.'
 DEFAULT_TIMEOUT = 7200  # 2 hours in seconds
 FORCE_KILL_DELAY = 5
@@ -34,6 +37,11 @@ def log_warn(message: str):
     sys.stderr.write(f"WARN: {message}\n")
 
 
+def log_info(message: str):
+    """输出信息到 stderr"""
+    sys.stderr.write(f"INFO: {message}\n")
+
+
 def resolve_timeout() -> int:
     """解析超时配置（秒）"""
     raw = os.environ.get('CODEX_TIMEOUT', '')
@@ -76,20 +84,72 @@ def parse_args():
             'mode': 'resume',
             'session_id': sys.argv[2],
             'task': sys.argv[3],
-            'model': sys.argv[4] if len(sys.argv) > 4 else DEFAULT_MODEL,
+            'workdir': sys.argv[4] if len(sys.argv) > 4 else DEFAULT_WORKDIR
-            'workdir': sys.argv[5] if len(sys.argv) > 5 else DEFAULT_WORKDIR
         }
     else:
         return {
             'mode': 'new',
             'task': sys.argv[1],
-            'model': sys.argv[2] if len(sys.argv) > 2 else DEFAULT_MODEL,
+            'workdir': sys.argv[2] if len(sys.argv) > 2 else DEFAULT_WORKDIR
-            'workdir': sys.argv[3] if len(sys.argv) > 3 else DEFAULT_WORKDIR
         }
 
 
-def build_codex_args(params: dict) -> list:
+def read_piped_task() -> Optional[str]:
-    """构建 codex CLI 参数"""
+    """
+    从 stdin 读取任务文本：
+    - 如果 stdin 是管道（非 tty）且存在内容，返回读取到的字符串
+    - 否则返回 None
+    """
+    import select
+
+    stdin = sys.stdin
+    if stdin is None or stdin.isatty():
+        log_info("Stdin is tty or None, skipping pipe read")
+        return None
+
+    # 使用 select 检查是否有数据可读（0 秒超时，非阻塞）
+    readable, _, _ = select.select([stdin], [], [], 0)
+    if not readable:
+        log_info("No data available on stdin")
+        return None
+
+    log_info("Reading from stdin pipe...")
+    data = stdin.read()
+    if not data:
+        log_info("Stdin pipe returned empty data")
+        return None
+
+    log_info(f"Read {len(data)} bytes from stdin pipe")
+    return data
+
+
+def should_stream_via_stdin(task_text: str, piped: bool) -> bool:
+    """
+    判定是否通过 stdin 传递任务：
+    - 有管道输入
+    - 文本包含换行
+    - 文本包含反斜杠
+    - 文本长度 > 800
+    """
+    if piped:
+        return True
+    if '\n' in task_text:
+        return True
+    if '\\' in task_text:
+        return True
+    if len(task_text) > 800:
+        return True
+    return False
+
+
+def build_codex_args(params: dict, target_arg: str) -> list:
+    """
+    构建 codex CLI 参数
+
+    Args:
+        params: 参数字典
+        target_arg: 最终传递给 codex 的参数（'-' 或具体 task 文本）
+    """
     if params['mode'] == 'resume':
         return [
             'codex', 'e',
@@ -97,39 +157,59 @@ def build_codex_args(params: dict) -> list:
             '--json',
             'resume',
             params['session_id'],
-            params['task']
+            target_arg
         ]
     else:
-        return [
+        base_args = [
             'codex', 'e',
-            '-m', params['model'],
+            '-m', DEFAULT_MODEL,
             '--dangerously-bypass-approvals-and-sandbox',
             '--skip-git-repo-check',
             '-C', params['workdir'],
             '--json',
-            params['task']
+            target_arg
         ]
 
+        return base_args
 
-def main():
-    params = parse_args()
-    codex_args = build_codex_args(params)
-    timeout_sec = resolve_timeout()
 
+def run_codex_process(codex_args, task_text: str, use_stdin: bool, timeout_sec: int):
+    """
+    启动 codex 子进程，处理 stdin / JSON 行输出和错误，成功时返回 (last_agent_message, thread_id)。
+    失败路径上负责日志和退出码。
+    """
     thread_id: Optional[str] = None
     last_agent_message: Optional[str] = None
+    process: Optional[subprocess.Popen] = None
 
     try:
-        # 启动 codex 子进程
+        # 启动 codex 子进程（文本模式管道）
+        log_info(f"Starting codex with args: {' '.join(codex_args[:5])}...")
         process = subprocess.Popen(
             codex_args,
+            stdin=subprocess.PIPE if use_stdin else None,
             stdout=subprocess.PIPE,
-            stderr=sys.stderr,  # 错误直接透传到 stderr
+            stderr=sys.stderr,
             text=True,
-            bufsize=1  # 行缓冲
+            bufsize=1,
         )
+        log_info(f"Process started with PID: {process.pid}")
+
+        # 如果使用 stdin 模式，写入任务到 stdin 并关闭
+        if use_stdin and process.stdin is not None:
+            log_info(f"Writing {len(task_text)} chars to stdin...")
+            process.stdin.write(task_text)
+            process.stdin.flush()  # 强制刷新缓冲区，避免大任务死锁
+            process.stdin.close()
+            log_info("Stdin closed")
 
         # 逐行解析 JSON 输出
+        if process.stdout is None:
+            log_error('Codex stdout pipe not available')
+            sys.exit(1)
+
+        log_info("Reading stdout...")
+
         for line in process.stdout:
             line = line.strip()
             if not line:
@@ -152,35 +232,26 @@ def main():
             except json.JSONDecodeError:
                 log_warn(f"Failed to parse line: {line}")
 
-        # 等待进程结束
+        # 等待进程结束并检查退出码
         returncode = process.wait(timeout=timeout_sec)
+        if returncode != 0:
+            log_error(f'Codex exited with status {returncode}')
+            sys.exit(returncode)
 
-        # 优先检查是否有有效输出，而非退出码
+        if not last_agent_message:
-        if last_agent_message:
+            log_error('Codex completed without agent_message output')
-            # 输出 agent_message
+            sys.exit(1)
-            sys.stdout.write(f"{last_agent_message}\n")
 
-            # 输出 session_id（如果存在）
+        return last_agent_message, thread_id
-            if thread_id:
-                sys.stdout.write(f"\n---\nSESSION_ID: {thread_id}\n")
-
-            # 有输出但退出码非零，输出警告而非失败
-            if returncode != 0:
-                log_warn(f'Codex completed with non-zero status {returncode} but produced valid output')
-
-            sys.exit(0)
-        else:
-            # 没有输出才算真正失败
-            log_error(f'Codex exited with status {returncode} without agent_message output')
-            sys.exit(returncode if returncode != 0 else 1)
 
     except subprocess.TimeoutExpired:
         log_error('Codex execution timeout')
-        process.kill()
+        if process is not None:
-        try:
+            process.kill()
-            process.wait(timeout=FORCE_KILL_DELAY)
+            try:
-        except subprocess.TimeoutExpired:
+                process.wait(timeout=FORCE_KILL_DELAY)
-            pass
+            except subprocess.TimeoutExpired:
+                pass
         sys.exit(124)
 
     except FileNotFoundError:
@@ -188,13 +259,64 @@ def main():
         sys.exit(127)
 
     except KeyboardInterrupt:
-        process.terminate()
+        log_error("Codex interrupted by user")
-        try:
+        if process is not None:
-            process.wait(timeout=FORCE_KILL_DELAY)
+            process.terminate()
-        except subprocess.TimeoutExpired:
+            try:
-            process.kill()
+                process.wait(timeout=FORCE_KILL_DELAY)
+            except subprocess.TimeoutExpired:
+                process.kill()
         sys.exit(130)
 
 
+def main():
+    log_info("Script started")
+    params = parse_args()
+    log_info(f"Parsed args: mode={params['mode']}, task_len={len(params['task'])}")
+    timeout_sec = resolve_timeout()
+    log_info(f"Timeout: {timeout_sec}s")
+
+    piped_task = read_piped_task()
+    piped = piped_task is not None
+    task_text = piped_task if piped else params['task']
+
+    use_stdin = should_stream_via_stdin(task_text, piped)
+
+    if use_stdin:
+        reasons = []
+        if piped:
+            reasons.append('piped input')
+        if '\n' in task_text:
+            reasons.append('newline')
+        if '\\' in task_text:
+            reasons.append('backslash')
+        if len(task_text) > 800:
+            reasons.append('length>800')
+
+        if reasons:
+            log_warn(f"Using stdin mode for task due to: {', '.join(reasons)}")
+
+    target_arg = '-' if use_stdin else params['task']
+    codex_args = build_codex_args(params, target_arg)
+
+    log_info('codex running...')
+
+    last_agent_message, thread_id = run_codex_process(
+        codex_args=codex_args,
+        task_text=task_text,
+        use_stdin=use_stdin,
+        timeout_sec=timeout_sec,
+    )
+
+    # 输出 agent_message
+    sys.stdout.write(f"{last_agent_message}\n")
+
+    # 输出 session_id（如果存在）
+    if thread_id:
+        sys.stdout.write(f"\n---\nSESSION_ID: {thread_id}\n")
+
+    sys.exit(0)
+
+
 if __name__ == '__main__':
     main()

skills/gemini/SKILL.md

@@ -0,0 +1,120 @@
+---
+name: gemini
+description: Execute Gemini CLI for AI-powered code analysis and generation. Use when you need to leverage Google's Gemini models for complex reasoning tasks.
+---
+
+# Gemini CLI Integration
+
+## Overview
+
+Execute Gemini CLI commands with support for multiple models and flexible prompt input. Integrates Google's Gemini AI models into Claude Code workflows.
+
+## When to Use
+
+- Complex reasoning tasks requiring advanced AI capabilities
+- Code generation and analysis with Gemini models
+- Tasks requiring Google's latest AI technology
+- Alternative perspective on code problems
+
+## Usage
+**Mandatory**: Run via uv with fixed timeout 7200000ms (foreground):
+```bash
+uv run ~/.claude/skills/gemini/scripts/gemini.py "<prompt>" [working_dir]
+```
+
+**Optional** (direct execution or using Python):
+```bash
+~/.claude/skills/gemini/scripts/gemini.py "<prompt>" [working_dir]
+# or
+python3 ~/.claude/skills/gemini/scripts/gemini.py "<prompt>" [working_dir]
+```
+
+## Environment Variables
+
+- **GEMINI_MODEL**: Configure model (default: `gemini-3-pro-preview`)
+  - Example: `export GEMINI_MODEL=gemini-3`
+
+## Timeout Control
+
+- **Fixed**: 7200000 milliseconds (2 hours), immutable
+- **Bash tool**: Always set `timeout: 7200000` for double protection
+
+### Parameters
+
+- `prompt` (required): Task prompt or question
+- `working_dir` (optional): Working directory (default: current directory)
+
+### Return Format
+
+Plain text output from Gemini:
+
+```text
+Model response text here...
+```
+
+Error format (stderr):
+
+```text
+ERROR: Error message
+```
+
+### Invocation Pattern
+
+When calling via Bash tool, always include the timeout parameter:
+
+```yaml
+Bash tool parameters:
+- command: uv run ~/.claude/skills/gemini/scripts/gemini.py "<prompt>"
+- timeout: 7200000
+- description: <brief description of the task>
+```
+
+Alternatives:
+
+```yaml
+# Direct execution (simplest)
+- command: ~/.claude/skills/gemini/scripts/gemini.py "<prompt>"
+
+# Using python3
+- command: python3 ~/.claude/skills/gemini/scripts/gemini.py "<prompt>"
+```
+
+### Examples
+
+**Basic query:**
+
+```bash
+uv run ~/.claude/skills/gemini/scripts/gemini.py "explain quantum computing"
+# timeout: 7200000
+```
+
+**Code analysis:**
+
+```bash
+uv run ~/.claude/skills/gemini/scripts/gemini.py "review this code for security issues: $(cat app.py)"
+# timeout: 7200000
+```
+
+**With specific working directory:**
+
+```bash
+uv run ~/.claude/skills/gemini/scripts/gemini.py "analyze project structure" "/path/to/project"
+# timeout: 7200000
+```
+
+**Using python3 directly (alternative):**
+
+```bash
+python3 ~/.claude/skills/gemini/scripts/gemini.py "your prompt here"
+```
+
+## Notes
+
+- **Recommended**: Use `uv run` for automatic Python environment management (requires uv installed)
+- **Alternative**: Direct execution `./gemini.py` (uses system Python via shebang)
+- Python implementation using standard library (zero dependencies)
+- Cross-platform compatible (Windows/macOS/Linux)
+- PEP 723 compliant (inline script metadata)
+- Requires Gemini CLI installed and authenticated
+- Supports all Gemini model variants (configure via `GEMINI_MODEL` environment variable)
+- Output is streamed directly from Gemini CLI

skills/gemini/scripts/gemini.py

@@ -0,0 +1,140 @@
+#!/usr/bin/env python3
+# /// script
+# requires-python = ">=3.8"
+# dependencies = []
+# ///
+"""
+Gemini CLI wrapper with cross-platform support.
+
+Usage:
+    uv run gemini.py "<prompt>" [workdir]
+    python3 gemini.py "<prompt>"
+    ./gemini.py "your prompt"
+"""
+import subprocess
+import sys
+import os
+
+DEFAULT_MODEL = os.environ.get('GEMINI_MODEL', 'gemini-3-pro-preview')
+DEFAULT_WORKDIR = '.'
+TIMEOUT_MS = 7_200_000  # 固定 2 小时，毫秒
+DEFAULT_TIMEOUT = TIMEOUT_MS // 1000
+FORCE_KILL_DELAY = 5
+
+
+def log_error(message: str):
+    """输出错误信息到 stderr"""
+    sys.stderr.write(f"ERROR: {message}\n")
+
+
+def log_warn(message: str):
+    """输出警告信息到 stderr"""
+    sys.stderr.write(f"WARN: {message}\n")
+
+
+def log_info(message: str):
+    """输出信息到 stderr"""
+    sys.stderr.write(f"INFO: {message}\n")
+
+
+def parse_args():
+    """解析位置参数"""
+    if len(sys.argv) < 2:
+        log_error('Prompt required')
+        sys.exit(1)
+
+    return {
+        'prompt': sys.argv[1],
+        'workdir': sys.argv[2] if len(sys.argv) > 2 else DEFAULT_WORKDIR
+    }
+
+
+def build_gemini_args(args) -> list:
+    """构建 gemini CLI 参数"""
+    return [
+        'gemini',
+        '-m', DEFAULT_MODEL,
+        '-p', args['prompt']
+    ]
+
+
+def main():
+    log_info('Script started')
+    args = parse_args()
+    log_info(f"Prompt length: {len(args['prompt'])}")
+    log_info(f"Working dir: {args['workdir']}")
+    gemini_args = build_gemini_args(args)
+    timeout_sec = DEFAULT_TIMEOUT
+    log_info(f"Timeout: {timeout_sec}s")
+
+    # 如果指定了工作目录，切换到该目录
+    if args['workdir'] != DEFAULT_WORKDIR:
+        try:
+            os.chdir(args['workdir'])
+        except FileNotFoundError:
+            log_error(f"Working directory not found: {args['workdir']}")
+            sys.exit(1)
+        except PermissionError:
+            log_error(f"Permission denied: {args['workdir']}")
+            sys.exit(1)
+        log_info('Changed working directory')
+
+    try:
+        log_info(f"Starting gemini with model {DEFAULT_MODEL}")
+        process = None
+        # 启动 gemini 子进程，直接透传 stdout 和 stderr
+        process = subprocess.Popen(
+            gemini_args,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            text=True,
+            bufsize=1  # 行缓冲
+        )
+
+        # 实时输出 stdout
+        for line in process.stdout:
+            sys.stdout.write(line)
+            sys.stdout.flush()
+
+        # 等待进程结束
+        returncode = process.wait(timeout=timeout_sec)
+
+        # 读取 stderr
+        stderr_output = process.stderr.read()
+        if stderr_output:
+            sys.stderr.write(stderr_output)
+
+        # 检查退出码
+        if returncode != 0:
+            log_error(f'Gemini exited with status {returncode}')
+            sys.exit(returncode)
+
+        sys.exit(0)
+
+    except subprocess.TimeoutExpired:
+        log_error(f'Gemini execution timeout ({timeout_sec}s)')
+        if process is not None:
+            process.kill()
+            try:
+                process.wait(timeout=FORCE_KILL_DELAY)
+            except subprocess.TimeoutExpired:
+                pass
+        sys.exit(124)
+
+    except FileNotFoundError:
+        log_error("gemini command not found in PATH")
+        log_error("Please install Gemini CLI: https://github.com/google/generative-ai-python")
+        sys.exit(127)
+
+    except KeyboardInterrupt:
+        if process is not None:
+            process.terminate()
+            try:
+                process.wait(timeout=FORCE_KILL_DELAY)
+            except subprocess.TimeoutExpired:
+                process.kill()
+        sys.exit(130)
+
+
+if __name__ == '__main__':
+    main()