Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/jzl-maker/EA-SKILLWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/jzl-maker/ea-skill/flash)<a href="https://agentmods.dev/commands/jzl-maker/ea-skill/flash"><img src="https://agentmods.dev/badge/commands/jzl-maker/ea-skill/flash/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/jzl-maker/ea-skill/flash"><img src="https://agentmods.dev/badge/commands/jzl-maker/ea-skill/flash.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.02266 |
| Opus 5.5 | $0.00000 | $0.00906 |
| Sonnet 5.5 | $0.00000 | $0.00453 |
| Haiku 4.5 | $0.00000 | $0.00227 |
Grade A, and why
flash scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
命令: /ea flash (烧录)
功能
烧录固件到 MCU。两条通道：
| 通道 | 工具 | 适用 |
|---|---|---|
--backend openocd（默认） |
OpenOCD | ST-Link / CMSIS-DAP / DAPLink |
--backend jlink-native |
JLink.exe | J-Link 探针——Windows 上装了 SEGGER 驱动时，OpenOCD 用不了 J-Link（见下） |
先探测环境
# OpenOCD 通道
python <SKILL>/tools/flash-openocd/scripts/openocd_flasher.py --detect
# J-Link 通道（确认 JLink.exe 在位）
python <SKILL>/tools/jlink-debug/scripts/jlink_debug.py --detect
调用方式
OpenOCD 通道（ST-Link / CMSIS-DAP / DAPLink）：
python <SKILL>/tools/flash-openocd/scripts/openocd_flasher.py \
--artifact <产物路径> \
--interface stlink \
--target target/stm32f4x.cfg
J-Link 原生通道（--backend jlink-native，烧录后独立回读比对）：
python <SKILL>/tools/flash-openocd/scripts/openocd_flasher.py \
--backend jlink-native \
--artifact build/app.hex \
--device N32G4FRRE
# 只回读校验（不烧），或看将要执行什么
python <SKILL>/tools/flash-openocd/scripts/openocd_flasher.py \
--backend jlink-native --artifact build/app.hex --verify-only
.bin 必须给 --base-address 0x08000000（BIN 自身不含地址信息，工具拒绝猜）。
.hex / .elf / .axf 自带地址，无需。加 --json 输出机读结论。
参数来源
| 参数 | 来源 |
|---|---|
--artifact |
从 build-keil 的产物路径获取（AXF/ELF/HEX） |
--interface |
从 --detect 探测结果获得（stlink/jlink/cmsis-dap） |
--target |
根据芯片型号选择（GD32F407 → target/stm32f4x.cfg） |
--backend |
探针是 J-Link → jlink-native；其余 → openocd |
--device |
jlink-native 专用，芯片型号（默认 N32G4FRRE） |
--settle |
jlink-native 专用，编程后静置秒数（默认 10）。别调 0，除非你确定板子不需要（见「编程后不能立刻回读」） |
结果处理
| 字段 | 检查要点 |
|---|---|
| 烧录状态 | success / failure |
| 校验状态 | verified / skipped |
| 失败分类 | connection-failure / target-response-abnormal / project-config-error |
自动决策
- 烧录成功 → 提示下一步（串口观察启动日志）
- 烧录失败 → 根据失败分类引导排查
常见错误
| 错误类型 | 原因 | 解决方案 |
|---|---|---|
| connection-failure | 调试器未连接或驱动问题 | 检查调试器连接、USB驱动 |
| target-response-abnormal | 芯片未进入调试模式 | 检查芯片供电、复位电路 |
| project-config-error | target 配置文件与芯片不匹配 | 检查 --target 参数 |
| Unsupported transport | --interface 参数错误 | 使用正确的接口类型 (stlink/jlink/cmsis-dap) |
LIBUSB_ERROR_NOT_SUPPORTED + No J-Link device found |
SEGGER 驱动占着 USB，OpenOCD 的 libusb 后端拿不到句柄 | 换 --backend jlink-native。不要试 transport select swd —— 见下 |
Verification failed @ 0x08000000（J-Link） |
多为误报（回读与产物逐字节一致） | 用 --backend jlink-native，工具会自己给结论，不要手工比 MD5 |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed 16efe2944e3b
- 1mo ago Changed · +78 lines 86aaa1c05551
- 1mo ago Changed · +16 lines c1d70a30bb9d
- 1mo ago First seen · 63 lines · 0 tokens per session scan A e248784be630
flash is a command published in the GitHub repository jzl-maker/EA-SKILL (9 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,266 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
rtos
Design a FreeRTOS or Zephyr task structure with sized IPC, a priority inversion walk, stack budgets, and failure-mode timing.
cortex-m
Generate Cortex-M startup code or a linker script, configure a core subsystem, decode a fault, or add DWT timing.
debug-embedded
Attach a debugger, halt and inspect a target, decode fault registers, or configure ITM, SWO, or ETM trace.
embedded-linux
Generate a Yocto or Buildroot config, flashing commands, a kernel module skeleton, or on-target debug commands.
esp-teach
One-time project setup -- discover hardware, find datasheets, persist context to CLAUDE.md.
tapeout-check
Run the tapeout precheck gate on a design before submission.