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.
npx skills add HorizonRobotics/OE-Skills --skill j6-board-monitorgit clone --depth 1 https://github.com/HorizonRobotics/OE-SkillsWrote 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/skills/horizonrobotics/oe-skills/j6-board-monitor)<a href="https://agentmods.dev/skills/horizonrobotics/oe-skills/j6-board-monitor"><img src="https://agentmods.dev/badge/skills/horizonrobotics/oe-skills/j6-board-monitor/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/skills/horizonrobotics/oe-skills/j6-board-monitor"><img src="https://agentmods.dev/badge/skills/horizonrobotics/oe-skills/j6-board-monitor.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 | $0.00216 | $0.07666 |
| Opus 5 | $0.00108 | $0.03833 |
| Sonnet 5 | $0.00043 | $0.01533 |
| Haiku 4.5 | $0.00022 | $0.00767 |
Grade C, and why
j6-board-monitor scanned grade C with 1 finding 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 12d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
ssh root@<IP> "rm -rf ${BOARD_WORKDIR}/*.hbm ${BOARD_WORKDIR}/*.log ${BOARD_WORKDIR}/*.csv ${BOARD_WORKDIR}/*.txt ${BOARD_WORKDIR}/*.sh ${BOARD_WORKDIR}/hrt_model_exec*" How it starts
The opening of the file, as written. The whole thing — 544 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Board Monitor
通过 SSH 在 J6 开发板上执行 BPU 占用率、DDR 带宽、内存使用的实时监控。支持三种场景:
- Scenario A:CV 模型在受控帧率推理期间同步采集资源数据(使用
hrt_model_exec) - Scenario B:独立监控板端硬件资源(无推理负载)
- Scenario C:LLM 模型在循环推理期间同步采集资源数据(使用
simple_demo_request)
适用场景
- 在指定帧率(如 10Hz)推理时监控 BPU 占用率、DDR 带宽、内存使用
- 独立监控板端硬件资源(无推理负载)
- LLM 模型板端推理期间的 BPU/DDR/内存监控(循环推理 + 同步采集)
- 评估模型在实车设计帧率下的资源消耗
- 对比不同模型/配置的板端资源占用
触发关键词:BPU 监控、DDR 带宽、内存使用、资源监控、设定帧率、10Hz 推理、LLM 推理监控、LLM 资源、simple_demo_request、hrt_ucp_monitor、hrut_ddr
部署前预检:部署模型到板端前,先阅读
board-preflight.md检查 ION 内存容量、L2M 配置和模型-板端兼容性。
工作流程
严格按照以下 8 步顺序执行,不可跳步。
前置条件
开始工作流之前,确认以下条件满足:
.horizon/.env.board必须存在:读取该文件获取BOARD_IP、BOARD_TYPE、BOARD_WORKDIR。如果文件不存在或不完整,先触发board-detectionskill 完成板卡检测,再继续。- SSH 免密登录:本 skill 的所有 SSH/SCP 命令假设已配置免密登录(密钥认证)。如果 SSH 连接要求输入密码,提示用户先配置密钥认证或提供密码,然后在所有
ssh/scp命令中通过sshpass -p <password>前缀传递。 - OE 工具链可用:
hrt_ucp_monitor、hrut_ddr、hrt_model_exec须已部署到板端。Step 2 会验证这一点。
Step 1:收集信息
从 .env.board 和用户输入收集以下信息。优先从 .env.board 读取,仅在文件中缺少时才询问用户。
| 项目 | 必填 | 默认值/来源 | 说明 |
|---|---|---|---|
| 板端 IP | 是 | .env.board → BOARD_IP |
优先从文件读取 |
| SSH 用户名 | 是 | root |
|
| 认证方式 | 是 | 密钥(免密) | 见前置条件第 2 条 |
| 板端工作目录 | 是 | .env.board → BOARD_WORKDIR |
优先从文件读取 |
| 模型文件路径 | Scenario A 必填 | - | 板端路径或本地路径(本地路径需 SCP 上传) |
| 目标帧率 (FPS) | Scenario A 必填 | - | 如 10Hz、30Hz |
| 监控时长 (秒) | 否 | 30 | 采集持续时间 |
| 监控目标 | 否 | bpu,ddr,memory | 选择监控项 |
场景自动检测:
- 用户提供了 CV 模型(.hbm 较小、有明确 FPS 需求) → Scenario A(受控推理 + 同步监控)
- 用户仅提供监控需求 → Scenario B(独立监控,无推理负载)
- 用户提供了 LLM/VLM 模型(.hbm 较大、无 FPS 控制需求) → Scenario C(循环推理 + 同步监控)
Scenario A vs C 判断依据:
- 模型文件 > 500MB 或用户明确提到 LLM/VLM → Scenario C
- 模型文件 < 500MB 且有明确帧率要求(如 10Hz) → Scenario A
Step 2:检查板卡连通性与平台
- 读取
.horizon/.env.board,获取BOARD_TYPE、BOARD_IP、BOARD_WORKDIR - SSH 连通性检查:
ssh -o ConnectTimeout=5 root@<IP> "echo ok" - 平台检测(决定 hrut_ddr 参数):
# 读取平台类型
BOARD_TYPE=$(grep '^BOARD_TYPE=' .horizon/.env.board | head -1 | cut -d= -f2)
if echo "$BOARD_TYPE" | grep -q "nash-p"; then
# J6P: 4 BPU cores, hrut_ddr 需要 per-core 参数
DDR_TYPE="bpu_p0"
BPU_CORES=4
else
# J6E: 1 BPU core, hrut_ddr 使用统一参数
DDR_TYPE="bpu"
BPU_CORES=1
fi
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 544 lines · 216 tokens per session scan C 65b9fc4c4681
j6-board-monitor is a skill published in the GitHub repository HorizonRobotics/OE-Skills (19 stars, last pushed 2mo ago), licensed Apache-2.0. It adds 216 tokens to every session and 7,666 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
gke-compute-classes
Configures, optimizes, and troubleshoots GKE ComputeClasses. Use when configuring Spot VMs with on-demand fallback, targeting specific accelerators (GPUs/TPUs) or machine families, restricting ComputeClass access, or debugging pending pods related to node pool auto-creation. Do not use for cluster-level Node Auto…
jetson-diagnostic
Read-only Jetson health snapshot for identity, memory, GPU, thermal, power, storage, services, and top processes.
doca-socket-relay
Use this skill when the operator is driving the DOCA Socket Relay to bridge a socket-oriented host application onto a BlueField DPU peer without rewriting it — picking the deployment shape (in-process, sidecar, or BlueField service container), configuring the host-side socket and the DPU-side forwarding endpoint…
offensive-z-wave
Z-Wave attack methodology — sniffing with Z-Force / EZ-Wave / RTL-SDR + ZniffMobile, S0 (legacy) network-key derivation flaw and key reuse, S2 (modern) ECDH commissioning analysis, replay/injection on unauthenticated nodes, default-key brute-force on test deployments, and home-automation hub pivots. Use when targeting…
hsb-flash
Flash the FPGA on an HSB board connected to an NVIDIA devkit. Supports HSB Lattice boards (FPGA versions 2407, 2412, 2507, 2510) and Leopard Imaging VB1940 "all-in-one" cameras (FPGA versions 2507, 2510). Uses release-specific YAML manifests and board-type-specific program commands. Lattice and VB1940 commands must…
jetson-validate-image
Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.