idea-mcp

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

IntelliJ IDEA MCP Skill

MCP is self-describing - this skill only documents known pitfalls and best practices not covered by MCP descriptions.

MCP 是自描述的——本Skill仅记录MCP描述未涵盖的已知陷阱和最佳实践。

🚨 Critical Pitfalls (MCP Doesn't Warn You)

🚨 关键陷阱（MCP 未发出警告）

Connection Requirements

连接要求

IDEA MUST be running with MCP server connected
Commands fail silently if IDEA is not running - no clear error message

必须运行IDEA且已连接MCP服务器
如果IDEA未运行，命令会静默失败——没有明确的错误提示

Path Confusion (Most Common Error)

路径混淆（最常见错误）

```
projectPath
```
: ABSOLUTE PATH to project root (e.g.,
```
/Users/name/project
```
)
- ❌
```
project
```
  (wrong - relative)
- ✅
```
/Users/name/project
```
  (correct - absolute)
```
filePath
```
/
```
pathInProject
```
/
```
directoryPath
```
: RELATIVE to project root

```
projectPath
```
：项目根目录的绝对路径（例如：
```
/Users/name/project
```
）
- ❌
```
project
```
  （错误 - 相对路径）
- ✅
```
/Users/name/project
```
  （正确 - 绝对路径）
```
filePath
```
/
```
pathInProject
```
/
```
directoryPath
```
：相对于项目根目录的路径

Search Command Schema Bugs

搜索命令 Schema 漏洞

MCP claims parameters are optional but they're REQUIRED:

search_in_files_by_text

search_in_files_by_regex

: MUST include

maxUsageCount

Without this → schema error
Start with
```
maxUsageCount: 10
```

```
find_files_by_name_keyword
```
&
```
find_files_by_glob
```
: MUST include
```
fileCountLimit
```
- Without this → schema error "data must have required property 'probablyHasMoreMatchingFiles'"
- Critical Bug: When
```
fileCountLimit > actual matching files
```
  , MCP server omits
```
probablyHasMoreMatchingFiles
```
  field → schema error
- Workaround: Start with
```
fileCountLimit: 1
```
  or
```
2
```
  , then increase if needed
- For common keywords (many matches), higher limits work fine
- If schema error, reduce
```
fileCountLimit
```
  (not increase)

MCP 声称部分参数是可选的，但实际上它们是必填的：

```
search_in_files_by_text
```
和
```
search_in_files_by_regex
```
：必须包含
```
maxUsageCount
```
参数
- 缺少该参数会导致Schema错误
- 建议初始值设为
```
maxUsageCount: 10
```
```
find_files_by_name_keyword
```
和
```
find_files_by_glob
```
：必须包含
```
fileCountLimit
```
参数
- 缺少该参数会导致Schema错误："data must have required property 'probablyHasMoreMatchingFiles'"
- 严重漏洞：当
```
fileCountLimit
```
  大于实际匹配文件数时，MCP服务器会省略
```
probablyHasMoreMatchingFiles
```
  字段，进而引发Schema错误
- 解决方法：初始值设为
```
fileCountLimit: 1
```
  或
```
2
```
  ，必要时再调高
- 对于常见关键词（匹配结果多），可以使用更高的限制值
- 如果出现Schema错误，降低
```
fileCountLimit
```
  （不要调高）

Rename Refactoring Limitations

重命名重构限制

⚠️ MCP doesn't document what rename_refactoring CANNOT do:

✅ Renames: Field/variable declarations and their references
❌ May have limitations with: Method parameters in signatures, certain symbol types (test before relying)
Always use
```
search_in_files_by_text
```
first to understand scope
Always verify with
```
get_file_problems
```
after renaming

⚠️ MCP 未说明 rename_refactoring 无法完成的操作：

✅ 可重命名：字段/变量声明及其引用
❌ 可能存在限制的场景：方法签名中的参数、某些符号类型（使用前请测试）
务必先使用
```
search_in_files_by_text
```
了解影响范围
务必在重命名后使用
```
get_file_problems
```
验证

High-Risk Operations

高风险操作

MCP doesn't emphasize danger levels:

```
execute_terminal_command
```
: HIGH RISK - Can run ANY shell command with full system permissions (no sandboxing). Use with extreme caution; verify commands before execution.
```
rename_refactoring
```
&
```
replace_text_in_file
```
: Modifies code globally
```
create_new_file
```
: Creates files on disk immediately
```
reformat_file
```
: Modifies formatting without confirmation

MCP 未强调操作的危险等级：

```
execute_terminal_command
```
：高风险——可以以完整系统权限运行任意Shell命令（无沙箱限制）。使用时需格外谨慎；执行前务必验证命令。
```
rename_refactoring
```
和
```
replace_text_in_file
```
：会全局修改代码
```
create_new_file
```
：会立即在磁盘上创建文件
```
reformat_file
```
：无需确认即可修改代码格式

✅ Essential Best Practices

✅ 核心最佳实践

Path discipline: Absolute for
```
projectPath
```
, relative for everything else
Count parameters: Always include
```
maxUsageCount
```
for search,
```
fileCountLimit
```
for file search
File search strategy: Start with
```
fileCountLimit: 1-2
```
; if schema error, reduce the limit (bug: limit > matches causes error)
Before refactoring: Search first to understand scope
After changes: Use
```
get_file_problems
```
to verify no errors introduced
Java accuracy: Prefer IDEA's
```
get_file_problems
```
over LSP (avoids false positives)
Text replacement: Get exact text with
```
get_file_text_by_path
```
before replacing
Test execution: Use
```
execute_run_configuration
```
for proper TeamCity-formatted output
Timeouts: Build/test may need 120-300s, not default 30s

路径规范：
```
projectPath
```
使用绝对路径，其他路径均使用相对路径
必填计数参数：搜索命令务必包含
```
maxUsageCount
```
，文件搜索命令务必包含
```
fileCountLimit
```
文件搜索策略：初始值设为
```
fileCountLimit: 1-2
```
；如果出现Schema错误，降低限制值（漏洞：限制值大于匹配数会引发错误）
重构前：先搜索以了解影响范围
修改后：使用
```
get_file_problems
```
验证未引入新错误
Java代码准确性：优先使用IDEA的
```
get_file_problems
```
而非LSP（避免误报）
文本替换：替换前先使用
```
get_file_text_by_path
```
获取准确文本
测试执行：使用
```
execute_run_configuration
```
以获取标准TeamCity格式的输出
超时设置：构建/测试可能需要120-300秒，而非默认的30秒

🔄 Common Workflows (Dongting Project)

🔄 常见工作流（洞庭项目）

Note: Examples use project path
/Users/huangli/dt/dongting
. Adjust
projectPath
for other projects.

注意：示例使用项目路径
/Users/huangli/dt/dongting
。其他项目请调整
projectPath
。

1. Find a Java Class by Name

1. 通过名称查找Java类

Use case: Locate

KvImpl.java

in the codebase.

yaml

Tool: IntelliJ-IDEA_find_files_by_name_keyword
Parameters:
  projectPath: "/Users/huangli/dt/dongting"  # Absolute path
  nameKeyword: "KvImpl"
  fileCountLimit: 2  # Start small; bug: limit > matches causes schema error

If no results: Try

find_files_by_glob

with pattern:

yaml

Tool: IntelliJ-IDEA_find_files_by_glob
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  globPattern: "**/*KvImpl*.java"
  fileCountLimit: 5

使用场景：定位代码库中的

KvImpl.java

。

yaml

Tool: IntelliJ-IDEA_find_files_by_name_keyword
Parameters:
  projectPath: "/Users/huangli/dt/dongting"  # Absolute path
  nameKeyword: "KvImpl"
  fileCountLimit: 2  # Start small; bug: limit > matches causes schema error

如果无结果：尝试使用

find_files_by_glob

并指定匹配模式：

yaml

Tool: IntelliJ-IDEA_find_files_by_glob
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  globPattern: "**/*KvImpl*.java"
  fileCountLimit: 5

2. Search for Text Across Codebase

2. 在代码库中搜索文本

Use case: Find all occurrences of "DtLog.getLogger".

yaml

Tool: IntelliJ-IDEA_search_in_files_by_text
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  searchText: "DtLog.getLogger"
  maxUsageCount: 20  # REQUIRED parameter
  caseSensitive: true

For regex patterns (e.g., find all log declarations):

yaml

Tool: IntelliJ-IDEA_search_in_files_by_regex
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  regexPattern: "private static final DtLog log = DtLogs\\.getLogger\\(.*\\)"
  maxUsageCount: 10

使用场景：查找所有 "DtLog.getLogger" 的出现位置。

yaml

Tool: IntelliJ-IDEA_search_in_files_by_text
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  searchText: "DtLog.getLogger"
  maxUsageCount: 20  # REQUIRED parameter
  caseSensitive: true

正则表达式搜索（例如：查找所有日志声明）：

yaml

Tool: IntelliJ-IDEA_search_in_files_by_regex
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  regexPattern: "private static final DtLog log = DtLogs\\.getLogger\\(.*\\)"
  maxUsageCount: 10

3. Run a Test via IDEA Run Configuration

3. 通过IDEA运行配置执行测试

Use case: Execute the "client:test" run configuration.

yaml

Tool: IntelliJ-IDEA_execute_run_configuration
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  configurationName: "client:test"
  timeout: 120000  # 2 minutes for tests

First, list available run configurations:

yaml

Tool: IntelliJ-IDEA_get_run_configurations
Parameters:
  projectPath: "/Users/huangli/dt/dongting"

使用场景：执行 "client:test" 运行配置。

yaml

Tool: IntelliJ-IDEA_execute_run_configuration
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  configurationName: "client:test"
  timeout: 120000  # 2 minutes for tests

先列出可用的运行配置：

yaml

Tool: IntelliJ-IDEA_get_run_configurations
Parameters:
  projectPath: "/Users/huangli/dt/dongting"

4. Check Compilation Errors in a File

4. 检查文件中的编译错误

Use case: Verify

server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java

has no errors.

yaml

Tool: IntelliJ-IDEA_get_file_problems
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  filePath: "server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java"
  errorsOnly: true

Note: More accurate than LSP for this multi-module Maven project.

使用场景：验证

server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java

无错误。

yaml

Tool: IntelliJ-IDEA_get_file_problems
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  filePath: "server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java"
  errorsOnly: true

注意：对于这个多模块Maven项目，该方法比LSP更准确。

5. Safe Refactoring: Rename a Field

5. 安全重构：重命名字段

Pre-verification: Search for current usage:

yaml

Tool: IntelliJ-IDEA_search_in_files_by_text
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  searchText: "currentTerm"
  maxUsageCount: 50

Execute rename:

yaml

Tool: IntelliJ-IDEA_rename_refactoring
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  pathInProject: "server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java"
  symbolName: "currentTerm"
  newName: "currentElectionTerm"

Post-verification: Check for errors:

yaml

Tool: IntelliJ-IDEA_get_file_problems
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  filePath: "server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java"

预验证：搜索当前字段的使用情况：

yaml

Tool: IntelliJ-IDEA_search_in_files_by_text
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  searchText: "currentTerm"
  maxUsageCount: 50

执行重命名：

yaml

Tool: IntelliJ-IDEA_rename_refactoring
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  pathInProject: "server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java"
  symbolName: "currentTerm"
  newName: "currentElectionTerm"

后验证：检查是否存在错误：

yaml

Tool: IntelliJ-IDEA_get_file_problems
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  filePath: "server/src/main/java/com/github/dtprj/dongting/raft/server/RaftNode.java"

6. Explore Project Structure

6. 探索项目结构

Use case: View module layout.

yaml

Tool: IntelliJ-IDEA_list_directory_tree
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  directoryPath: "."  # Project root
  maxDepth: 3

To see all open files (helpful for context):

yaml

Tool: IntelliJ-IDEA_get_all_open_file_paths
Parameters:
  projectPath: "/Users/huangli/dt/dongting"

使用场景：查看模块布局。

yaml

Tool: IntelliJ-IDEA_list_directory_tree
Parameters:
  projectPath: "/Users/huangli/dt/dongting"
  directoryPath: "."  # Project root
  maxDepth: 3

查看所有已打开文件（有助于了解上下文）：

yaml

Tool: IntelliJ-IDEA_get_all_open_file_paths
Parameters:
  projectPath: "/Users/huangli/dt/dongting"

🤔 When to Use: IDEA MCP vs LSP vs Bash Tools

🤔 工具选择：IDEA MCP vs LSP vs Bash工具

Decision Matrix

决策矩阵

Task	Recommended Tool	Why
Java code analysis (errors, warnings)	IDEA MCP ( `get_file_problems` )	More accurate for multi-module Maven projects; avoids LSP false positives
Running tests	IDEA MCP ( `execute_run_configuration` )	Proper test environment, TeamCity-formatted output, dependency resolution
Refactoring (rename, move)	IDEA MCP ( `rename_refactoring` )	Semantic understanding, cross-file updates, safer than manual edits
Search code (text, regex)	IDEA MCP ( `search_in_files_by_text` )	Fast, indexes entire project; better than `grep` for large codebases
Find files (by name, glob)	IDEA MCP ( `find_files_by_*` )	Fast indexed search; use workarounds for schema bugs
Read file content	LSP ( `read` tool) or IDEA MCP ( `get_file_text_by_path` )	LSP for quick reads; IDEA for large files or binary detection
Build/compile	Bash ( `mvn` commands)	Full control, reproducible; IDEA build may have caching issues
Git operations	Bash ( `git` commands)	Standard, predictable; IDEA MCP lacks git tools
Project exploration	IDEA MCP ( `list_directory_tree` )	Visual tree structure; better than `ls` for nested directories
Terminal commands	IDEA MCP ( `execute_terminal_command` ) or Bash	IDEA for integrated terminal; Bash for complex pipelines

任务	推荐工具	原因
Java代码分析（错误、警告）	IDEA MCP ( `get_file_problems` )	对多模块Maven项目更准确；避免LSP误报
运行测试	IDEA MCP ( `execute_run_configuration` )	提供合适的测试环境、TeamCity格式输出、依赖解析
重构（重命名、移动）	IDEA MCP ( `rename_refactoring` )	具备语义理解能力、跨文件更新、比手动编辑更安全
代码搜索（文本、正则）	IDEA MCP ( `search_in_files_by_text` )	速度快，索引整个代码库；比 `grep` 更适合大型代码库
查找文件（按名称、通配符）	IDEA MCP ( `find_files_by_*` )	基于索引的快速搜索；针对Schema漏洞使用解决方法
读取文件内容	LSP（ `read` 工具）或 IDEA MCP ( `get_file_text_by_path` )	LSP适合快速读取；IDEA适合大文件或二进制文件检测
构建/编译	Bash（ `mvn` 命令）	完全可控、可复现；IDEA构建可能存在缓存问题
Git操作	Bash（ `git` 命令）	标准、可预测；IDEA MCP缺少Git工具
项目探索	IDEA MCP ( `list_directory_tree` )	可视化树状结构；比 `ls` 更适合嵌套目录
终端命令	IDEA MCP ( `execute_terminal_command` ) 或 Bash	IDEA适合集成终端；Bash适合复杂管道操作

Key Principles

核心原则

IDEA MCP when: Need IDE intelligence (refactoring, accurate errors, test execution)
LSP when: Quick file reads, symbol navigation (if IDEA not available)
Bash when: Builds, Git, system operations, or when IDEA MCP has schema bugs

Remember: IDEA must be running for MCP tools to work. If IDEA is unavailable, fall back to LSP/Bash.

使用IDEA MCP的场景：需要IDE智能能力（重构、精准错误检测、测试执行）
使用LSP的场景：快速读取文件、符号导航（当IDEA不可用时）
使用Bash的场景：构建、Git操作、系统操作，或IDEA MCP存在Schema漏洞时

注意：使用MCP工具时必须运行IDEA。如果IDEA不可用， fallback到LSP或Bash工具。

⏱️ Performance Considerations & Timeouts

⏱️ 性能考量与超时设置

Recommended Timeouts by Operation

各操作推荐超时时间

Operation	Recommended Timeout	Notes
File search ( `find_files_by_*` )	30000 (30s)	Usually fast; start with `fileCountLimit: 1-2` to avoid schema bug
Text search ( `search_in_files_by_*` )	60000 (60s)	Can be slower for large codebases; use `maxUsageCount`
Read file ( `get_file_text_by_path` )	10000 (10s)	Fast for text files; binary files may fail
Get file problems ( `get_file_problems` )	15000 (15s)	Quick analysis
Run configuration ( `execute_run_configuration` )	300000 (5min)	Tests/builds can take minutes; adjust based on test size
Build project ( `IntelliJ-IDEA_build_project` )	300000 (5min)	Full rebuild may be slow
Refactoring ( `rename_refactoring` )	60000 (60s)	Usually fast for single symbols
Directory listing ( `list_directory_tree` )	20000 (20s)	Depends on project size

操作	推荐超时时间	说明
文件搜索 ( `find_files_by_*` )	30000（30秒）	通常速度快；初始值设为 `fileCountLimit: 1-2` 以避免Schema漏洞
文本搜索 ( `search_in_files_by_*` )	60000（60秒）	大型代码库可能较慢；使用 `maxUsageCount` 限制结果数
读取文件 ( `get_file_text_by_path` )	10000（10秒）	文本文件读取速度快；二进制文件可能读取失败
获取文件问题 ( `get_file_problems` )	15000（15秒）	分析速度快
运行配置 ( `execute_run_configuration` )	300000（5分钟）	测试/构建可能需要数分钟；根据测试规模调整
构建项目 ( `IntelliJ-IDEA_build_project` )	300000（5分钟）	完整重构可能较慢
重构 ( `rename_refactoring` )	60000（60秒）	单个符号重构通常速度快
目录列表 ( `list_directory_tree` )	20000（20秒）	取决于项目规模

Performance Tips

性能优化技巧

Limit search results: Use
```
maxUsageCount
```
and
```
fileCountLimit
```
to avoid excessive data transfer
Cache project state: IDEA MCP maintains project index; repeated searches are faster
Avoid redundant calls: Use
```
get_all_open_file_paths
```
to see what's already loaded
Batch operations: Combine searches with careful planning (but MCP is stateless)
Fallback for large operations: For full project builds, use Bash
```
mvn
```
commands instead

限制搜索结果：使用
```
maxUsageCount
```
和
```
fileCountLimit
```
避免过多数据传输
缓存项目状态：IDEA MCP维护项目索引；重复搜索速度更快
避免冗余调用：使用
```
get_all_open_file_paths
```
查看已加载的文件
批量操作：通过周密规划组合搜索操作（但MCP是无状态的）
大型操作的备选方案：全项目构建使用Bash的
```
mvn
```
命令

Memory & Resource Considerations

内存与资源考量

IDEA MCP runs within IDEA's JVM; large operations may affect IDE performance
Searching huge files (>10MB) may cause timeouts; use LSP
```
read
```
with offset/limit instead
Concurrent MCP requests may queue; avoid parallel operations on same project

IDEA MCP运行在IDEA的JVM中；大型操作可能影响IDE性能
搜索超大文件（>10MB）可能超时；改用LSP的
```
read
```
工具并指定偏移量/限制
同一项目的并发MCP请求可能排队；避免并行操作

🔧 Quick Troubleshooting

🔧 快速故障排除

Common Error Messages and Solutions

常见错误信息与解决方案

Error Message	Likely Cause	Solution
`Structured content does not match the tool's output schema: data must have required property 'probablyHasMoreMatchingFiles'`	Missing `fileCountLimit` or `maxUsageCount` parameter	Add required parameter: `fileCountLimit: 1` (for file search) or `maxUsageCount: 10` (for text search)
`MCP error -32602: Structured content does not match the tool's output schema`	`fileCountLimit` > actual matches (MCP bug)	Reduce `fileCountLimit` to 1 or 2; common keywords can use higher limits
`Command failed: Connection refused`	IDEA not running or MCP server not connected	Verify IDEA is running with MCP plugin active; wait for connection
`Command timed out after X milliseconds`	Operation taking longer than default timeout	Increase `timeout` parameter (e.g., 120000 for tests, 30000 for builds)
`File not found` or `Path does not exist`	Incorrect path (absolute vs relative confusion)	Use absolute path for `projectPath` , relative path for `filePath`
`No occurrences found`	Search text not found or path incorrect	Verify search text exists; check `projectPath` and file paths
`Could not get document` or `File is binary`	Trying to read binary/ non-text file	Use `get_file_text_by_path` only for text files; check file type
`Project directory not found`	`projectPath` incorrect or IDEA project not loaded	Verify project is fully loaded in IDEA; use absolute path

错误信息	可能原因	解决方案
`Structured content does not match the tool's output schema: data must have required property 'probablyHasMoreMatchingFiles'`	缺少 `fileCountLimit` 或 `maxUsageCount` 参数	添加必填参数：文件搜索设 `fileCountLimit: 1` ，文本搜索设 `maxUsageCount: 10`
`MCP error -32602: Structured content does not match the tool's output schema`	`fileCountLimit` 大于实际匹配数（MCP漏洞）	降低 `fileCountLimit` 至1或2；常见关键词可使用更高值
`Command failed: Connection refused`	IDEA未运行或MCP服务器未连接	验证IDEA已运行且MCP插件已激活；等待连接建立
`Command timed out after X milliseconds`	操作耗时超过默认超时时间	调高 `timeout` 参数（例如：测试设120000，构建设30000）
`File not found` 或 `Path does not exist`	路径错误（绝对/相对路径混淆）	`projectPath` 使用绝对路径， `filePath` 使用相对路径
`No occurrences found`	搜索文本不存在或路径错误	验证搜索文本存在；检查 `projectPath` 和文件路径
`Could not get document` 或 `File is binary`	尝试读取二进制/非文本文件	`get_file_text_by_path` 仅用于文本文件；检查文件类型
`Project directory not found`	`projectPath` 错误或IDEA项目未加载	验证项目已在IDEA中完全加载；使用绝对路径

"Structured content does not match" (schema error)

"Structured content does not match"（Schema错误）

Add missing
```
maxUsageCount
```
or
```
fileCountLimit
```
parameter
For
find_files_by_name_keyword
:
```
fileCountLimit
```
> actual matches causes this bug. Reduce the limit to 1 or 2, then increase if needed

添加缺失的
```
maxUsageCount
```
或
```
fileCountLimit
```
参数
针对
find_files_by_name_keyword
：当
```
fileCountLimit
```
大于实际匹配数时会触发该漏洞。降低限制值至1或2，必要时再调高

Commands fail/timeout

命令失败/超时

Verify IDEA is running with MCP server connected
Check MCP plugin is active in IDEA settings
Wait for connection (may take seconds)
Verify project is fully loaded in IDEA

验证IDEA已运行且已连接MCP服务器
检查IDEA设置中MCP插件已激活
等待连接建立（可能需要数秒）
验证项目已在IDEA中完全加载

Search returns no results

搜索无结果

Verify
```
projectPath
```
is absolute and correct
Verify file paths are relative to project root
Use
```
list_directory_tree
```
to verify structure exists

验证
```
projectPath
```
是正确的绝对路径
验证文件路径是相对于项目根目录的相对路径
使用
```
list_directory_tree
```
验证结构存在

File operation fails

文件操作失败

For
```
create_new_file
```
: Use
```
overwrite: true
```
if file exists

For

replace_text_in_file

: Get exact text with

get_file_text_by_path

first

Check file is not locked by another process

对于
```
create_new_file
```
：如果文件已存在，使用
```
overwrite: true
```
参数
对于
```
replace_text_in_file
```
：替换前先使用
```
get_file_text_by_path
```
获取准确文本
检查文件是否被其他进程锁定