Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
App基础指标+智能巡检(异动报告)
v1.0.0查询友盟 (UMeng) 应用统计数据分析，支持通过 APPKEY 获取应用的基础指标信息如新增用户数、活跃用户数等。当用户提到"友盟"、"umeng"、"APPKEY"、"新增用户"、"活跃用户"或需要查询应用统计数据时使用此技能。
⭐ 0· 41·0 current·0 all-time
byUmeng+@squall0925
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description match the included code: the repository contains a full Python AOP/SDK for Umeng with requests to Umeng endpoints (e.g., gateway.open.umeng.com and mobile.umeng.com) and helper scripts for querying metrics and outlier reports. That aligns with the stated purpose. However the skill metadata lists no required environment variables or primary credential even though the SDK and SKILL.md clearly require API credentials (apiKey / apiSecurity).
Instruction Scope
SKILL.md clearly instructs the agent to load credentials from a config file (./umeng-config.json, ~/umeng-config.json, or skill-dir) or environment variables (UMENG_API_KEY, UMENG_API_SECURITY) and then call Umeng APIs. Those instructions are narrowly scoped to the stated functionality. A minor inconsistency: the README/UPDATE_LOG claims aop.__init__ was updated to auto-load config, but the provided aop/__init__.py is a lightweight implementation that does not implement an automatic _load_umeng_config function — the outlier helper imports umeng_config.py directly to get credentials. The skill will read files from the current directory, user home, or the skill directory (as documented) — expected for a credential-loading helper, but this behaviour increases the surface where credentials can be read, so users should be aware.
Install Mechanism
No install specification is declared (instruction-only for installation), so nothing will be downloaded or executed automatically by an installer. The repository includes many Python files (SDK + utility scripts) but no external installer or remote-downloads. This lowers install-time risk. The package does contain packaging and validation scripts (scripts/package_skill.py, quick_validate.py) which are not required for runtime but are not themselves harmful.
Credentials
The skill requires Umeng API credentials (apiKey / apiSecurity) and documents environment variables UMENG_API_KEY / UMENG_API_SECURITY and a config file, but the skill metadata declares no required env vars or primary credential. That mismatch is problematic because automated systems (and users) may not be warned that secrets are needed. The code will read credentials from CWD, home directory, or the skill directory — legitimate for this API client, but the skill requesting access to filesystem locations (home, cwd) should have been reflected in the metadata.
Persistence & Privilege
The skill does not request always:true and does not modify other skills or system-wide configuration. It contains only code to call Umeng APIs and helper utilities; there is no evidence it attempts to persist itself beyond its directory or to access other skills' credentials. Autonomous invocation is allowed by default (disable-model-invocation: false) which is normal; this combined with the credential handling is why you should carefully control which agents get these credentials.
What to consider before installing
This skill is an Umeng (友盟) Python SDK and legitimately needs your Umeng apiKey and apiSecurity. However: (1) the skill's metadata does not declare those required credentials even though SKILL.md and the code expect them — treat this as a red flag and do not supply secrets without inspecting the code; (2) the code will look for a config file in the current directory, your home directory, or the skill directory — check those locations for accidental credential leakage and do not store long-lived secrets in shared repos; (3) review umeng_config.py and umeng_get_outlier_points.py to confirm credentials are only used to call Umeng endpoints (gateway.open.umeng.com / mobile.umeng.com) and not exfiltrated elsewhere; (4) prefer running the code in a restricted/sandboxed environment and set file permissions (chmod 600) on any config file; (5) ask the publisher to update the skill metadata to list required env vars (UMENG_API_KEY, UMENG_API_SECURITY) and clarify automatic loading behavior — if they do and code audit is clean, the mismatches here would be resolved and the skill would be coherent.Like a lobster shell, security has layers — review code before you run it.
latestvk979fs0pd712h2bgkhm3hrgm7s84fk6m
License
MIT-0
Free to use, modify, and redistribute. No attribution required.