65/100 (Medium)

OpenClaw： suspicious
VirusTotal： benign
StaticScan： clean

OpenClaw： suspicious

The skill appears to do what it claims (extract frames and build dataset.jsonl) and contains only local file operations, but there are incoherences and omissions — e.g., missing declared binary requir...

[内容已截断]

VirusTotal： benign VT 报告

静态扫描： clean

No suspicious patterns detected.

README 未提供

{
    "latestVersion": {
        "_creationTime": 1773814381176,
        "_id": "k97f7bes39vmaxqtw0p0htej6h834b0x",
        "changelog": "# Changelog — hair-cam-anno\n\n所有格式基于 [Keep a Changelog](https:\/\/keepachangelog.com\/zh-CN\/1.1.0\/)，\n版本号遵循 [语义化版本](https:\/\/semver.org\/lang\/zh-CN\/)。\n\n---\n\n## [0.1.0] — 2026-03-18\n\n### 新增\n- 初始版本发布\n- **视频帧提取脚本** `scripts\/extract_frames.py`\n  - 基于 ffmpeg，支持 mp4\/avi\/mkv\/mov\/flv\/wmv\/webm 格式\n  - 可配置提取帧率（默认 0.5 fps）和每视频最大帧数（默认 4 帧）\n  - 自动生成 `manifest.json`（含视频元信息：分辨率、时长、帧数、编码格式）\n- **dataset.jsonl 构建脚本** `scripts\/build_jsonl.py`\n  - 读取 `annotations.json`，生成符合 VL 模型微调规范的 `dataset.jsonl`\n  - 内置自动验证：字段完整性、标签合法性、风险等级合法性、simple_description 长度（≤20字）、description 长度（≥50字）\n- **System Prompt 模板** `references\/system-prompt.md`\n  - 定义 VL 模型角色、6 步分析流程、JSON 输出格式约束\n  - 包含 15 个预定义场景标签（system_suggest_0 ~ system_suggest_14）\n  - 风险等级：none \/ low \/ medium \/ high\n- **标签与风险等级参考** `references\/labels-reference.md`\n  - 15 个标签定义与含义说明\n  - 4 级风险等级定义及典型场景\n  - 数据集类别要求（物体\/行为动作\/环境）数量标准\n- **SKILL.md** 技能主文件\n  - 三步工作流程：帧提取 → 逐视频分析标注 → 汇总生成 dataset.jsonl\n  - 标签选择规则指导",
        "changelogSource": "user",
        "createdAt": 1773814381176,
        "version": "1.0.0"
    },
    "owner": {
        "_creationTime": 0,
        "_id": "publishers:missing",
        "displayName": "Aowind",
        "handle": "aowind",
        "image": "https:\/\/avatars.githubusercontent.com\/u\/26450413?v=4",
        "kind": "user",
        "linkedUserId": "kn789707vk375dmrgyd8cy9drx8322cn"
    },
    "ownerHandle": "aowind",
    "skill": {
        "_creationTime": 1773814381176,
        "_id": "kd787w30be3cj9d91m9z0hvzcn835wbe",
        "badges": [],
        "createdAt": 1773814381176,
        "displayName": "sjht-cam-anno",
        "latestVersionId": "k97f7bes39vmaxqtw0p0htej6h834b0x",
        "ownerUserId": "kn789707vk375dmrgyd8cy9drx8322cn",
        "slug": "sjht-cam-anno",
        "stats": {
            "comments": 0,
            "downloads": 39,
            "installsAllTime": 0,
            "installsCurrent": 0,
            "stars": 1,
            "versions": 1
        },
        "summary": "安防摄像头视频 VL 模型微调数据集标注工具。用于从安防摄像头视频中提取关键帧、分析视频内容、生成结构化标注（含环境\/人物\/行为\/风险描述），并输出符合 dataset.jsonl 格式的微调训练数据。Use when 用户需要对安防摄像头视频进行数据标注、生成 VL 模型训练数据集、处理 \/root\/hair-...",
        "tags": {
            "latest": "k97f7bes39vmaxqtw0p0htej6h834b0x"
        },
        "updatedAt": 1774336905929
    }
}