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clear审计通过

bullet-tracer子弹追踪器

Agent Skill

bullet-tracer 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

294

周安装

12

GitHub Stars

1

下载量

95
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

3

许可证

MIT

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:bullet-tracer(子弹追踪器)
来源仓库:https://github.com/sebastiaanwouters/dotagents
仓库路径:skills/bullet-tracer
安装命令:
npx skills add https://github.com/sebastiaanwouters/dotagents --skill bullet-tracer
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/sebastiaanwouters/dotagents --skill bullet-tracer

简介

基于《程序员修炼之道》的子弹追踪法,构建端到端最小可行垂直切片。

  • 适合快速验证需求可行性,覆盖 UI、API、逻辑、数据库和测试全链路。
  • 不适用于纯重构或单层修改类任务,需选择性应用核心思想。
  • 使用时应评估项目兼容性,避免在不适用场景强行套用模式。
  • bullet-tracer 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Bullet Tracer Skill

From The Pragmatic Programmer: "Use Tracer Bullets to Find the Target" — build a minimal end-to-end vertical slice that touches all layers first, then expand.

Compatibility Note

Not all tickets/features are fully compatible with the tracer bullet approach (e.g., pure refactors, config changes, single-layer fixes). Be smart: apply tracer bullet concepts to compatible parts, use general best practices for the rest.

Philosophy

  • Tracer bullet!= prototype: You keep the code, it becomes the foundation
  • Vertical slice: Touch ALL layers (UI → API → logic → DB → tests) in one thin path
  • Fast feedback: Validate architecture and integration points early
  • Hardcoded is OK: Initial tracer can use minimal/fake data

Workflow

Phase 1: Tracer Bullet (MUST DO FIRST)

  1. Identify all layers the feature touches (e.g., frontend, API, service, DB, config)
  2. Implement the thinnest possible path through ALL layers:

- One happy path only - Hardcoded values acceptable - Skip edge cases, validation, error handling - Must be runnable/testable end-to-end

  1. Write one integration test that exercises the full path
  2. Verify it works — run the test, manually test if needed

Phase 2: Expand

Only after tracer works, expand systematically:

  1. Replace hardcoded values with real implementations
  2. Add validation and error handling
  3. Add edge cases and additional paths
  4. Add unit tests for individual components
  5. Refactor for production quality

Example

Feature: "User can save favorite items"

Tracer (Phase 1):

- Frontend: Add button that calls POST /api/favorites with hardcoded itemId
- API: Endpoint that inserts into favorites table
- DB: Create favorites table with user_id, item_id
- Test: Integration test that clicks button, verifies DB row exists

Expand (Phase 2):

  • Real item selection UI
  • Validation (item exists, not already favorited)
  • Error handling and user feedback
  • Unfavorite functionality
  • List favorites page
  • Unit tests for each layer

Rules

  1. NEVER skip the tracer phase — always build e2e first
  2. NEVER expand before tracer works — resist the urge to "do it properly"
  3. Keep tracer minimal — if you're adding "nice to haves", stop
  4. Tracer must touch ALL layers — if it doesn't, it's not a tracer
  5. Tracer must be testable — if you can't verify it works, add a test

Anti-patterns

❌ Building the perfect DB schema before any code ❌ Implementing full validation before happy path works ❌ Writing all unit tests before integration test ❌ Building frontend in isolation from backend ❌ "I'll connect it later"

Checklist

Before moving to Phase 2, verify:

  • Code touches every layer the feature needs
  • Can run/test the feature end-to-end
  • At least one integration test passes

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

能力 5

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

Claude Code

26.67%
按下载量换算25

windsurf

23.28%
按下载量换算22

trae

19.66%
按下载量换算19

OpenCode

13.33%
按下载量换算13

Codex

7.82%
按下载量换算7

github-copilot

3.82%
按下载量换算4

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。

来源信息

继续浏览同类 Skills