Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/HorizonRobotics/OE-Skillsnpx agentmods add skills/horizonrobotics/oe-skills/oe-llm-package-detectionWrote 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/oe-llm-package-detection)<a href="https://agentmods.dev/skills/horizonrobotics/oe-skills/oe-llm-package-detection"><img src="https://agentmods.dev/badge/skills/horizonrobotics/oe-skills/oe-llm-package-detection/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/oe-llm-package-detection"><img src="https://agentmods.dev/badge/skills/horizonrobotics/oe-skills/oe-llm-package-detection.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 | $0.00101 | $0.03272 |
| Opus 5.5 | $0.00040 | $0.01309 |
| Sonnet 5.5 | $0.00020 | $0.00654 |
| Haiku 4.5 | $0.00010 | $0.00327 |
Grade A, and why
oe-llm-package-detection 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 2mo 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 — 224 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OE-LLM 包环境检测
执行方式
本 Skill 应通过 subagent 执行。 主 agent 在前置检查中发现
.env.oe-llm-package缺失或不完整时,应将本文件的完整内容作为 subagent prompt 派发执行。subagent 完成后汇报写入结果(OE-LLM 路径、版本、执行模式),主 agent 读取.env.oe-llm-package继续后续流程。
目标
检测 OE-LLM 包路径、版本及本地环境匹配情况,写入 .horizon/.env.oe-llm-package,供后续所有 LLM 工具链任务直接使用。
触发条件
任何涉及 LLM 量化、LLM 压缩、LLM 编译、板端 LLM 推理、LLM 精度评估的任务进入 horizon-router 前,如果 .horizon/.env.oe-llm-package 不存在,顶层 Skill 会中断任务并提示用户:
未检测到 OE-LLM 包环境配置(
.horizon/.env.oe-llm-package)。请提供 OE-LLM 包路径,或回复"跳过"暂不配置。
- 用户提供路径 → 进入本检测流程
- 用户回复"跳过" → 记录跳过(仅当次对话有效,下次仍会提示),继续后续任务
LLM 任务的识别关键词:LLM、大语言模型、VLM、视觉语言模型、llm_compression、AWQ、GPTQ、RTN、SmoothQuant、LLM 量化、LLM 编译、LLM 推理、oellm、InternVL、Qwen-VL、LLaMA 等。
检测流程
1. 检查 .horizon/.env.oe-llm-package 是否存在
- 文件存在且内容完整(包含
OE_LLM_DIR、OE_LLM_VERSION、EXECUTION_MODE等字段)→ 直接读取,跳过后续步骤 - 文件不存在或不完整 → 进入步骤 2
2. 定位 OE-LLM 包路径
按优先级查找:
- 环境变量:
OE_LLM_DIR、OPEN_EXPLORER_LLM_DIR、HORIZON_OE_LLM_DIR - 项目配置文件:
.env、.horizon/oe-llm.env、CLAUDE.md中声明的路径 - 常见路径探测:
/open_explorer_llm、~/open_explorer_llm、/opt/openexplorer_llm,以及/mnt/oe-cli-test/下以horizon_j6_open_explorer_llm开头的目录 - 以上都没有 → 询问用户 OE-LLM 包路径
找到路径后,验证目录中存在 OE-LLM 包的标志性文件(以下至少两个):
run_docker.shllm_compression/目录runtime/目录
注意:OE-LLM 包没有标准 OE 包的 samples/、docs/、toolchain/ 目录,也没有 package/host/ai_toolchain/ 子目录。
3. 采集 OE-LLM 版本信息
从 OE-LLM 包目录中提取以下信息:
OE-LLM 包整体版本:
- 从目录名提取(如
horizon_j6_open_explorer_llm_v2.0.0_rc3-py310_20260615→v2.0.0_rc3) - 检查
README-CN、README-EN中的版本信息 - 以上都无法确定时询问用户
组件版本(从 llm_compression/deps_version.conf 解析):
| 字段 | 含义 | 示例值 |
|---|---|---|
HBDK_VERSION |
HBDK4 编译器版本 | 4.11.2 |
HORIZON_PLUGIN_PYTORCH_VERSION |
PyTorch 量化插件版本 | 3.3.4 |
HBM_INFER_VERSION |
HBM 推理引擎版本 | 3.15.3 |
LLM_COMPRESSION_VERSION |
LLM 压缩工具版本 | 2.0.2 |
TORCH_VERSION |
配套 PyTorch 版本(含 CUDA) | 2.8.0+cu128 |
PYTHON_VERSION |
配套 Python 版本 | py310 |
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.
- 2mo ago First seen · 224 lines · 101 tokens per session scan A a0ce72ea2725
oe-llm-package-detection is a skill published in the GitHub repository HorizonRobotics/OE-Skills (27 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 101 tokens to every session and 3,272 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-08-30.
oe-llm-package-detection compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| oe-llm-package-detectionHorizonRobotics/OE-Skills | Skill | 27 | 4mo ago | Apache-2.0 | A | 101 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 133k | yesterday | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 99k | yesterday | Apache-2.0 | A | 19 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | 4d ago | MIT | A | 200 | Claude Code, Codex |
| systematic-debuggingobra/superpowers | Skill | 313k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | yesterday | MIT | A | 103 | Claude Code |