Ai Intelligent Behavior Analysis

行为分析，动作识别 + 异常检测。

Security Scan

The skill's instructions tell you to download and run unreviewed code from GitHub (git clone + pip install && python app.py) while declaring no credentials, no install spec, and no runtime details — that mismatch and the implied remote execution are concerning.

Purpose & Capability

The skill claims real-time video analytics, alerts, and integrations (camera streams, push/alerting) but declares no required environment variables, endpoints, or binaries. Real deployment would normally require camera endpoints, API keys/webhook targets, storage or model weights, and possibly GPU/runtime requirements; those are absent, which is inconsistent with the stated capabilities.

Instruction Scope

SKILL.md instructs the operator to git clone a GitHub repo and run pip install -r requirements.txt and python app.py. That directs the agent/user to fetch and execute arbitrary remote code without any guardrails, configuration guidance, or explicit limits on data access — the runtime instructions therefore broaden scope beyond the simple description and could enable network access, file I/O, or data exfiltration depending on the remote code.

Install Mechanism

There is no packaged install spec in the registry; instead SKILL.md asks to clone a GitHub repo and run pip install on requirements.txt. While GitHub is a common host, cloning and executing unreviewed code plus installing arbitrary PyPI packages is a moderate-to-high risk install pattern because it pulls external artifacts that were not included in the skill bundle or statically scanned.

Credentials

No environment variables, credentials, or config paths are declared despite features that normally require credentials or endpoints (camera RTSP/HTTP URLs, push notification API keys, storage, or model hosting). The lack of declared secrets/configs is disproportionate and makes it unclear what sensitive data the remote app will request or transmit.

Persistence & Privilege

The skill is not marked always:true and does not request system-wide config changes in the provided files. Autonomous invocation is allowed by platform default, but there is no explicit persistent privilege requested in the skill metadata.

Scan Findings in Context

[NO_CODE_FILES_FOR_STATIC_SCAN] unexpected: The regex/static scanner had no code files to analyze (skill is instruction-only). This is not expected for a skill whose SKILL.md instructs cloning and executing a remote repository: the static scan couldn't evaluate the remote repo's contents, so risk remains in the external code that would be fetched.

What to consider before installing

Proceed cautiously. Do not run git clone && pip install && python app.py on production or a host with sensitive data. Before installing: inspect the referenced GitHub repository (verify owner, read README, check app.py and requirements.txt), review any network calls or environment-variable usage, look for hardcoded endpoints or secrets, and scan the repo for malicious packages. If you must test it, run it in an isolated VM or container with no access to sensitive networks or credentials. Ask the publisher for a homepage, a signed release, and documentation of required camera endpoints, alert webhook configuration, and data retention/policy details. If you need help auditing the external repo, provide its URL and I can review the repo contents (code, requirements, and network calls) for further analysis.

AI 智能行为分析系统

描述

行为分析，动作识别 + 异常检测。

功能

• 动作识别（动作分类）
• 异常行为（打架/跌倒）
• 人群分析（密度/流向）
• 轨迹分析（行为模式）
• 预警推送（实时告警）

定价

• 基础版：¥399/月（1 路/月）
• 专业版：¥1999/月（10 路/月）
• 企业版：¥5999/月（无限路）

适用场景

• 安防监控
• 工厂安全
• 养老监护
• 校园安全

技术栈

• Python + FastAPI
• 行为识别模型
• 视频分析
• 实时告警

安装

git clone https://github.com/openclaw-skills/ai-intelligent-behavior-analysis
cd ai-intelligent-behavior-analysis
pip install -r requirements.txt
python app.py

---

创建：2026-03-13 作者：OpenClaw Skills Team