CCM
/Skills
SkillsMCPMarketplacesDigestToolsAdvertise

This week in Claude

Every Monday: Claude Code, Agent SDK, MCP, and the Anthropic platform moves worth your time.

Skills by Category
Frontend DevelopmentBackend & APIsTesting & QASecurityDevOps & CI/CDGit & Pull RequestsDocumentationCode Review & QualityAI & Agent BuildingSkill Development
MCP Servers by Category
Sales & MarketingWeb & Browser AutomationDatabasesAI & LLM ToolsCloud & InfrastructureCommunication & MessagingDeveloper ToolsDesign & CreativeDocuments & KnowledgeSearch & Web Crawling
Marketplaces by Category
AI Agents & OrchestrationLLM IntegrationDevelopment ToolsFrontend & UIBackend & APIsDatabasesTesting & Code QualityDevOps & CloudSecurity & ComplianceGit & Version Control

Claude Code Marketplaces

Discover Claude Code plugins, extensions, and tools. Automatically updated directory of Anthropic Claude AI marketplaces with development tools, productivity plugins, and integrations.

Resources

  • Browse Skills
  • Browse MCP Servers
  • Browse Marketplaces
  • Skill index
  • MCP index
  • Marketplace index
  • Plugins Reference

Community

  • About
  • Tools
  • Feedback
  • Privacy Policy
  • Advertise

Built for the Claude Code community with Claude Code by mertbuilds.com

Independent project, not affiliated with Anthropic
zhinkgit avatar

Jlink

zhinkgit/embeddedskills
488 installs545 stars
Summary

If you're working with J-Link probes and need to flash firmware, read RTT logs, or debug ARM targets without leaving your editor, this handles the full workflow. It wraps JLink.exe, JLinkGDBServerCL, and arm-none-eabi-gdb into a unified interface with JSON output. You can halt execution, step through instructions, set breakpoints by address, or drop into source-level GDB sessions when you have an ELF file. The config setup is two-tier: tool paths live in the skill directory, device parameters (chip model, SWD vs JTAG, speed) live in your workspace. It won't guess your chip type or auto-select probes, which is the right call when a wrong guess bricks something.

Install to Claude Code

npx -y skills add zhinkgit/embeddedskills --skill jlink --agent claude-code

Installs into .claude/skills of the current project.

CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
Files
SKILL.mdView on GitHub

J-Link 下载与在线调试

本 skill 提供 J-Link 探针的设备探测、固件烧录、内存读写、寄存器查看、目标复位、RTT 日志读取,以及轻量在线调试和 GDB 源码级调试能力。

配置

环境级配置(skill/config.json)

skill 目录下的 config.json 包含环境级配置(工具路径、端口号等),首次使用前确认 exe 路径正确:

{
  "exe": "C:\\Program Files\\SEGGER\\JLink\\JLink.exe",
  "gdbserver_exe": "C:\\Program Files\\SEGGER\\JLink\\JLinkGDBServerCL.exe",
  "rtt_exe": "C:\\Program Files\\SEGGER\\JLink\\JLinkRTTClient.exe",
  "gdb_exe": "C:\\Program Files\\Arm\\GNU Toolchain mingw-w64-x86_64-arm-none-eabi\\bin\\arm-none-eabi-gdb.exe",
  "serial_no": "",
  "rtt_telnet_port": 0,
  "swo_command": [],
  "operation_mode": 1
}
  • exe:JLink.exe 完整路径(必填)
  • gdbserver_exe:JLinkGDBServerCL.exe 路径,RTT 和 GDB 调试需要
  • rtt_exe:JLinkRTTClient.exe 路径
  • gdb_exe:arm-none-eabi-gdb 路径,GDB 源码级调试需要
  • serial_no:默认探针序列号,多探针场景下使用
  • rtt_telnet_port:RTT 端口,0 表示使用工具默认值
  • swo_command:可选,完整 SWO viewer 命令数组,供 jlink_swo.py 包装
  • operation_mode:1 直接执行 / 2 输出风险摘要但不阻塞 / 3 执行前确认

工程级配置(.embeddedskills/config.json)

设备参数(device/interface/speed)统一在工作区的 .embeddedskills/config.json 中管理:

{
  "jlink": {
    "device": "STM32F407VG",
    "interface": "SWD",
    "speed": "4000"
  }
}
  • device:芯片型号(如 STM32F407VG、GD32F470ZG)
  • interface:调试接口,SWD 或 JTAG,默认 SWD
  • speed:调试速率 kHz,默认 4000

参数解析优先级:CLI 显式参数 > .embeddedskills/config.json(工程级)> skill/config.json(环境级)> .embeddedskills/state.json > 默认值/报错

成功执行后,确认过的 device/interface/speed 会自动写回工程配置。

子命令

基础操作

子命令用途风险
info探测探针与目标连通性低
flash烧录固件(.hex / .bin / .elf)高
read-mem读取内存区域低
write-mem写入内存高
regs查看 CPU 寄存器低
reset复位目标芯片高
rtt读取 RTT 日志输出低
swo包装外部 SWO viewer 为统一事件流低

在线调试(JLink Commander)

子命令用途风险
halt暂停 CPU,返回寄存器状态低
go恢复 CPU 运行低
step单步执行(支持指定步数),返回执行的指令和寄存器低
run-to设置断点并运行,等待命中后返回状态低

GDB 源码级调试

子命令用途依赖
gdb backtrace/locals查看调用栈和局部变量arm-none-eabi-gdb
gdb break/continue/next/step/finish/untilone-shot 控制执行流arm-none-eabi-gdb
gdb frame/print/watch/disassemble/threads/crash-reportone-shot 源码级诊断arm-none-eabi-gdb

执行流程

  1. 读取 skill/config.json,确认 exe 路径有效
  2. 读取 .embeddedskills/config.json 获取工程级配置(device/interface/speed)
  3. 读取 .embeddedskills/state.json 获取历史状态
  4. 参数解析优先级:CLI 显式参数 > .embeddedskills/config.json(工程级)> skill/config.json(环境级)> .embeddedskills/state.json > 默认值/报错
  5. 若当前动作需要 device 且仍为空,直接要求用户补充,绝不猜测
  6. 多探针场景未指定 serial_no 时,列出探针让用户选择,不自动选择
  7. 按 operation_mode 决定是否需要确认后执行
  8. 使用模板生成临时 .jlink 命令文件,调用 JLink.exe 时带 -NoGui 1 -ExitOnError 1 -AutoConnect 1
  9. 解析输出和返回码,返回结构化结果
  10. 成功执行后,将确认过的 device/interface/speed 写回 .embeddedskills/config.json

脚本调用

skill 目录下有四个 Python 脚本,使用标准库实现,无额外依赖。

jlink_exec.py — 基础操作 + 轻量调试

# 探测连通性
python <skill-dir>/scripts/jlink_exec.py info --device GD32F470ZG --json

# 烧录固件
python <skill-dir>/scripts/jlink_exec.py flash --file build/app.hex --device GD32F470ZG --json

# 烧录 .bin(必须提供地址)
python <skill-dir>/scripts/jlink_exec.py flash --file build/app.bin --device GD32F470ZG --address 0x08000000 --json

# 读取内存
python <skill-dir>/scripts/jlink_exec.py read-mem --address 0x08000000 --length 256 --device GD32F470ZG --json

# 写入内存
python <skill-dir>/scripts/jlink_exec.py write-mem --address 0x20000000 --value 0x12345678 --device GD32F470ZG --json

# 查看寄存器
python <skill-dir>/scripts/jlink_exec.py regs --device GD32F470ZG --json

# 复位目标
python <skill-dir>/scripts/jlink_exec.py reset --device GD32F470ZG --json

# 暂停 CPU
python <skill-dir>/scripts/jlink_exec.py halt --device GD32F470ZG --json

# 恢复运行
python <skill-dir>/scripts/jlink_exec.py go --device GD32F470ZG --json

# 单步执行(3 步)
python <skill-dir>/scripts/jlink_exec.py step --device GD32F470ZG --count 3 --json

# 运行到断点地址
python <skill-dir>/scripts/jlink_exec.py run-to --device GD32F470ZG --address 0x08001234 --timeout-ms 3000 --json

通用可选参数:--interface SWD|JTAG、--speed 4000、--serial-no <序列号>、--exe <JLink.exe路径>

jlink_rtt.py — RTT 日志读取

python <skill-dir>/scripts/jlink_rtt.py --device GD32F470ZG --json

可选参数:--serial-no、--channel、--encoding、--rtt-port、--gdbserver-exe <路径>、--rtt-exe <路径>

RTT 工作原理:脚本先通过 JLinkGDBServerCL.exe 建立调试连接,再启动 JLinkRTTClient.exe 读取 RTT 数据。--json 模式输出 JSON Lines。

jlink_swo.py — SWO 事件流包装

# 使用 config.json 里的 swo_command
python <skill-dir>/scripts/jlink_swo.py --json

# 或显式传入 viewer 命令
python <skill-dir>/scripts/jlink_swo.py \
  --viewer-cmd JLinkSWOViewerCL.exe -device GD32F470ZG -itf SWD -speed 4000 \
  --json

jlink_swo.py 不直接实现 SWO 协议,而是把外部 viewer 的 stdout/stderr 统一包装成 JSON Lines,便于上层 workflow 或 AI 继续消费。

jlink_gdb.py — GDB 源码级调试(需要 arm-none-eabi-gdb)

# 执行自定义 GDB 命令序列
python <skill-dir>/scripts/jlink_gdb.py run \
  --gdbserver-exe <路径> --gdb-exe <arm-none-eabi-gdb路径> \
  --device GD32F470ZG --elf build/app.elf \
  --commands "break main" "continue" "backtrace" "info locals" --json

# 快捷:获取调用栈
python <skill-dir>/scripts/jlink_gdb.py backtrace \
  --gdbserver-exe <路径> --gdb-exe <路径> \
  --device GD32F470ZG --elf build/app.elf --json

# 快捷:查看局部变量
python <skill-dir>/scripts/jlink_gdb.py locals \
  --gdbserver-exe <路径> --gdb-exe <路径> \
  --device GD32F470ZG --elf build/app.elf --json

GDB 调试需要 ELF 文件才能进行源码级调试(断点到函数名、查看变量)。没有 ELF 时仍可使用地址级调试。

输出格式

所有脚本以 JSON 格式返回,基础字段为 status(ok/error)、action、summary、details,并可能附带 context、artifacts、metrics、state、next_actions、timing。流式观测命令使用 JSON Lines,并统一输出 source、channel_type、stream_type。

成功示例:

{
  "status": "ok",
  "action": "halt",
  "summary": "已暂停,PC=0x08049ABC",
  "details": {
    "device": "GD32F470ZG",
    "registers": { "PC": "0x08049ABC", "R0": "0x00000004", "..." : "..." }
  }
}

step 示例(包含执行的指令):

{
  "status": "ok",
  "action": "step",
  "summary": "单步3次,PC=0x08049AB4",
  "details": {
    "steps": [
      { "address": "0x08049AB8", "opcode": "80 1B", "instruction": "SUBS R0, R0, R6" },
      { "address": "0x08049ABA", "opcode": "A8 42", "instruction": "CMP R0, R5" },
      { "address": "0x08049ABC", "opcode": "FA D3", "instruction": "BCC #-0x0C" }
    ],
    "registers": { "PC": "0x08049AB4", "..." : "..." }
  }
}

run-to 示例(断点命中):

{
  "status": "ok",
  "action": "run-to",
  "summary": "断点命中 @ 0x08049AB4,PC=0x08049AB4",
  "details": {
    "bp_address": "0x08049AB4",
    "bp_hit": true,
    "registers": { "PC": "0x08049AB4", "..." : "..." }
  }
}

核心规则

  • 不自动猜测 device 芯片型号,缺失时必须询问用户
  • 多探针场景不自动选择探针,必须让用户指定序列号
  • 参数解析优先级为:CLI 显式参数 > .embeddedskills/config.json(工程级)> skill/config.json(环境级)> .embeddedskills/state.json > 默认值/报错
  • .bin 文件必须显式提供烧录地址,缺失时报错
  • 连接失败时给出排查建议(检查连线、供电、接口类型、速度),不自动尝试更激进参数
  • 烧录、写内存、复位在参数完整且用户意图明确时直接执行
  • run-to 的断点在单次 JLink 会话内完成设置和清除,不存在跨会话 handle 问题
  • 结果回显中始终包含目标芯片、接口类型和执行动作
  • 产物路径(elf/file)不进入工程配置,仍只依赖 state.json
  • last_flash/last_debug 等运行状态继续写入 state.json

调试典型工作流

快速排查(JLink Commander)

halt → regs → read-mem → step → go

适合查看当前执行位置、寄存器状态、内存值,无需 ELF 和 GDB。

断点调试(JLink Commander)

run-to(address) → regs → read-mem → go

在指定地址设置断点并等待命中,查看此时的状态。

源码级调试(GDB)

gdb run --elf app.elf --commands "break main" "continue" "backtrace" "info locals"

需要 ELF 文件和 arm-none-eabi-gdb,支持函数名断点和变量查看。

参考

遇到芯片型号问题时可查阅 references/common_devices.md。

Featured
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
MCP-ready Email SendingMCP-ready Email Sending
MCP-ready Email Sending
Plug Mailtrap into your AI workflow and let it handle the email.
Connect Mailtrap MCP →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Capacitor - Shared memory for your team’s coding agents.
Capacitor - Shared memory for your team’s coding agents.
Make coding agent sessions - Searchable, Shareable, Vendor-neutral & Scored.
Try For Free →
CodeScene MCP ServerCodeScene MCP Server
CodeScene MCP Server
Your agent targets a perfect 10 Code Health score. Deterministic. Every commit.
Try For Free →
Give your AI the whole web as clean markdownGive your AI the whole web as clean markdown
Give your AI the whole web as clean markdown
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
inference shell
inference shell
create and run specialised agents in minutes
build now →
First SeenJun 3, 2026
View on GitHub

More from zhinkgit/embeddedskills

All 9 skills →
  • Gcc481
  • Openocd480
  • Workflow470
  • Net452
  • Probe Rs448
  • Keil647
  • Serial604
  • Can520

Recommended

juliusbrussee avatar
caveman

juliusbrussee/caveman

Ultra-compressed communication mode. Cuts output tokens 65% (measured) by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.
443.2k
98.9k
mattpocock avatar
grill-me

mattpocock/skills

A relentless interview to sharpen a plan or design.
896.5k
221k
shadcn avatar
improve

shadcn/improve

Survey any codebase as a senior advisor and produce prioritized, self-contained implementation plans for OTHER models/agents to execute. Strictly read-only on source code — never implements, fixes, or refactors anything itself. Use when asked to audit a codebase, find improvement opportunities (bugs, security, performance, test coverage, tech debt, migrations, DX), suggest features or where to take the project next (roadmap, product direction), or generate handoff plans for another agent to implement.
29.7k
8.9k
obra avatar
systematic-debugging

obra/superpowers

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
229.2k
273.5k
forrestchang avatar
karpathy-guidelines

forrestchang/andrej-karpathy-skills

Behavioral guidelines to reduce common LLM coding mistakes through explicit assumptions, simplicity, and verifiable success criteria.
17.9k
191.7k
vercel-labs avatar
find-skills

vercel-labs/skills

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.
3M
28.2k