Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计通过

jscodeshift-codemodsjscodeshift 代码模块

Agent Skill

jscodeshift-codemods 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

423

周安装

18

GitHub Stars

5

下载量

148
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:jscodeshift-codemods(jscodeshift 代码模块)
来源仓库:https://github.com/third774/dotfiles
仓库路径:skills/jscodeshift-codemods
安装命令:
npx skills add https://github.com/third774/dotfiles --skill jscodeshift-codemods
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/third774/dotfiles --skill jscodeshift-codemods

简介

用于查找、检索和筛选相关信息。jscodeshift-codemods 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

  • 适合根据关键词或任务场景快速定位候选结果。
  • 可辅助研究类任务的信息收集环节。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。
  • 建议结合具体需求验证搜索策略的有效性。
  • 需注意来源仓库的维护状态和实际功能边界。

SKILL.md

jscodeshift Codemods

Core Philosophy: Transform AST nodes, not text. Let recast handle printing to preserve formatting and structure.

When to Use

Use codemods for:

  • API migrations - Library upgrades (React Router v5→v6, enzyme→RTL)
  • Pattern standardization - Enforce coding conventions across codebase
  • Deprecation removal - Remove deprecated APIs systematically
  • Large-scale refactoring - Rename functions, restructure imports, update patterns

Don't use codemods for:

  • One-off changes (faster to do manually)
  • Changes requiring semantic understanding (business logic)
  • Non-deterministic transformations

Codemod Workflow

Copy this checklist and track your progress:

Codemod Progress:
- [ ] Phase 1: Identify Patterns
  - [ ] Collect before/after examples from real code
  - [ ] Document transformation rules
  - [ ] Identify edge cases
- [ ] Phase 2: Create Test Fixtures
  - [ ] Create input fixture with pattern to transform
  - [ ] Create expected output fixture
  - [ ] Verify test fails (TDD)
- [ ] Phase 3: Implement Transform
  - [ ] Find target nodes
  - [ ] Apply transformation
  - [ ] Return modified source
- [ ] Phase 4: Handle Edge Cases
  - [ ] Add fixtures for edge cases
  - [ ] Handle already-transformed code (idempotency)
  - [ ] Handle missing dependencies
- [ ] Phase 5: Validate at Scale
  - [ ] Dry run on target codebase
  - [ ] Review sample of changes
  - [ ] Run with --fail-on-error

Project Structure

Standard codemod project layout:

codemods/
├── my-transform.ts                    # Transform implementation
├── __tests__/
│   └── my-transform-test.ts           # Test file
└── __testfixtures__/
    ├── my-transform.input.ts          # Input fixture
    ├── my-transform.output.ts         # Expected output
    ├── edge-case.input.ts             # Additional fixtures
    └── edge-case.output.ts

Transform Module Anatomy

Every transform exports a function with this signature:

import type { API, FileInfo, Options } from "jscodeshift";

export default function transform(
  fileInfo: FileInfo,
  api: API,
  options: Options
): string | null | undefined {
  const j = api.jscodeshift;
  const root = j(fileInfo.source);

  // Find and transform nodes
  root
    .find(j.Identifier, { name: "oldName" })
    .forEach((path) => {
      path.node.name = "newName";
    });

  // Return transformed source, null to skip, or undefined for no change
  return root.toSource();
}

Return values:

ReturnMeaning
stringTransformed source code
nullSkip this file (no output)
undefinedNo changes made

Key objects:

ObjectPurpose
fileInfo.sourceOriginal file contents
fileInfo.pathFile path being transformed
api.jscodeshiftThe jscodeshift library (usually aliased as j)
api.statsCollect statistics during dry runs
api.reportPrint to stdout

Testing with defineTest

jscodeshift provides fixture-based testing utilities:

// __tests__/my-transform-test.ts
jest.autoMockOff();
const defineTest = require("jscodeshift/dist/testUtils").defineTest;

// Basic test - uses my-transform.input.ts → my-transform.output.ts
defineTest(__dirname, "my-transform");

// Named fixtures for edge cases
defineTest(__dirname, "my-transform", null, "already-transformed");
defineTest(__dirname, "my-transform", null, "missing-import");
defineTest(__dirname, "my-transform", null, "multiple-occurrences");

Fixture naming:

__testfixtures__/
├── my-transform.input.ts              # Default input
├── my-transform.output.ts             # Default output
├── already-transformed.input.ts       # Named fixture input
├── already-transformed.output.ts      # Named fixture output

Running tests:

# Run all codemod tests
npx jest codemods/__tests__/

# Run specific transform tests
npx jest codemods/__tests__/my-transform-test.ts

# Run with verbose output
npx jest codemods/__tests__/my-transform-test.ts --verbose

Collection API Quick Reference

The jscodeshift Collection API provides chainable methods:

MethodPurposeExample
find(type, filter?)Find nodes by typeroot.find(j.CallExpression, {callee: {name: 'foo'}})
filter(predicate)Filter collection.filter(path => path.node.arguments.length > 0)
forEach(callback)Iterate and mutate.forEach(path => {path.node.name = 'new'})
replaceWith(node)Replace matched nodes.replaceWith(j.identifier('newName'))
remove()Remove matched nodes.remove()
insertBefore(node)Insert before each match.insertBefore(j.importDeclaration(...))
insertAfter(node)Insert after each match.insertAfter(j.expressionStatement(...))
closest(type)Find nearest ancestor.closest(j.FunctionDeclaration)
get()Get first path.get()
paths()Get all paths as array.paths()
size()Count matches.size()

Chaining pattern:

root
  .find(j.CallExpression, { callee: { name: "oldFunction" } })
  .filter((path) => path.node.arguments.length === 2)
  .forEach((path) => {
    path.node.callee.name = "newFunction";
  });

Common Node Types

Node TypeRepresentsExample Code
IdentifierVariable/function namesfoo, myVar
CallExpressionFunction callsfoo(), obj.method()
MemberExpressionProperty accessobj.prop, arr[0]
ImportDeclarationImport statementsimport {x} from 'y'
ImportSpecifierNamed imports{x} in import
ImportDefaultSpecifierDefault importsx in import x from
VariableDeclarationVariable declarationsconst x = 1
VariableDeclaratorIndividual variablex = 1 part
FunctionDeclarationNamed functionsfunction foo() {}
ArrowFunctionExpressionArrow functions() => {}
ObjectExpressionObject literals{a: 1, b: 2}
ArrayExpressionArray literals[1, 2, 3]
LiteralPrimitive values'string', 42, true
StringLiteralString values'hello'

Common Transformation Patterns

Rename Import Source

// Change: import { x } from 'old-package'
// To:     import { x } from 'new-package'

root
  .find(j.ImportDeclaration, { source: { value: "old-package" } })
  .forEach((path) => {
    path.node.source.value = "new-package";
  });

Rename Named Import

// Change: import { oldName } from 'package'
// To:     import { newName } from 'package'

root
  .find(j.ImportSpecifier, { imported: { name: "oldName" } })
  .forEach((path) => {
    path.node.imported.name = "newName";
    // Also rename local if not aliased
    if (path.node.local.name === "oldName") {
      path.node.local.name = "newName";
    }
  });

Add Import If Missing

// Add: import { newThing } from 'package'

const existingImport = root.find(j.ImportDeclaration, {
  source: { value: "package" },
});

if (existingImport.size() === 0) {
  // Add new import at top of file
  const newImport = j.importDeclaration(
    [j.importSpecifier(j.identifier("newThing"))],
    j.literal("package")
  );

  root.find(j.Program).get("body", 0).insertBefore(newImport);
}

Rename Function Calls

// Change: oldFunction(arg)
// To:     newFunction(arg)

root
  .find(j.CallExpression, { callee: { name: "oldFunction" } })
  .forEach((path) => {
    path.node.callee.name = "newFunction";
  });

Transform Function Arguments

// Change: doThing(a, b, c)
// To:     doThing({ a, b, c })

root
  .find(j.CallExpression, { callee: { name: "doThing" } })
  .filter((path) => path.node.arguments.length === 3)
  .forEach((path) => {
    const [a, b, c] = path.node.arguments;
    path.node.arguments = [
      j.objectExpression([
        j.property("init", j.identifier("a"), a),
        j.property("init", j.identifier("b"), b),
        j.property("init", j.identifier("c"), c),
      ]),
    ];
  });

Track Variable Usage Across Scope

// Find what variable an import is bound to, then find all usages

root.find(j.ImportSpecifier, { imported: { name: "useHistory" } }).forEach((path) => {
  const localName = path.node.local.name; // Could be aliased

  // Find all calls using this variable
  root
    .find(j.CallExpression, { callee: { name: localName } })
    .forEach((callPath) => {
      // Transform each usage
    });
});

Replace Entire Expression

// Change: history.push('/path')
// To:     navigate('/path')

root
  .find(j.CallExpression, {
    callee: {
      type: "MemberExpression",
      object: { name: "history" },
      property: { name: "push" },
    },
  })
  .replaceWith((path) => {
    return j.callExpression(j.identifier("navigate"), path.node.arguments);
  });

Anti-Patterns

Over-Matching

// BAD: Matches ANY identifier named 'foo'
root.find(j.Identifier, { name: "foo" });

// GOOD: Match specific context (function calls named 'foo')
root.find(j.CallExpression, { callee: { name: "foo" } });

Ignoring Scope

// BAD: Assumes 'history' always means the router history
root.find(j.Identifier, { name: "history" });

// GOOD: Verify it came from the expected import
const historyImport = root.find(j.ImportSpecifier, {
  imported: { name: "useHistory" },
});
if (historyImport.size() === 0) return; // Skip file

Not Checking Idempotency

// BAD: Adds import every time, even if already present
root.find(j.Program).get("body", 0).insertBefore(newImport);

// GOOD: Check first
const existingImport = root.find(j.ImportDeclaration, {
  source: { value: "package" },
});
if (existingImport.size() === 0) {
  root.find(j.Program).get("body", 0).insertBefore(newImport);
}

Destructive Transforms

// BAD: Rebuilds node from scratch, loses comments and formatting
path.replace(
  j.callExpression(j.identifier("newFn"), [j.literal("arg")])
);

// GOOD: Mutate existing node to preserve metadata
path.node.callee.name = "newFn";

Testing Only Happy Path

// BAD: Only one test fixture
defineTest(__dirname, "my-transform");

// GOOD: Cover edge cases
defineTest(__dirname, "my-transform");
defineTest(__dirname, "my-transform", null, "already-transformed");
defineTest(__dirname, "my-transform", null, "aliased-import");
defineTest(__dirname, "my-transform", null, "no-matching-code");

Debugging Transforms

Dry Run with Print

# See output without writing files
npx jscodeshift -t my-transform.ts target/ --dry --print

Log Node Structure

root.find(j.CallExpression).forEach((path) => {
  console.log(JSON.stringify(path.node, null, 2));
});

Verbose Mode

# Show transformation stats
npx jscodeshift -t my-transform.ts target/ --verbose=2

Fail on Errors

# Exit with code 1 if any file fails
npx jscodeshift -t my-transform.ts target/ --fail-on-error

CLI Quick Reference

# Basic usage
npx jscodeshift -t transform.ts src/

# TypeScript/TSX files
npx jscodeshift -t transform.ts src/ --parser=tsx --extensions=ts,tsx

# Dry run (no changes)
npx jscodeshift -t transform.ts src/ --dry

# Print output to stdout
npx jscodeshift -t transform.ts src/ --print

# Limit parallelism
npx jscodeshift -t transform.ts src/ --cpus=4

# Ignore patterns
npx jscodeshift -t transform.ts src/ --ignore-pattern="**/*.test.ts"

Integration

Complementary skills:

  • writing-tests - For test-first codemod development
  • systematic-debugging - When transforms produce unexpected results
  • verification-before-completion - Verify codemod works before claiming done

Language-specific patterns:

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

34.13%
按下载量换算51

Claude

30.84%
按下载量换算46

Cursor

20.78%
按下载量换算31

Gemini CLI

10.61%
按下载量换算16

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills