web-data-extractor（网页数据采集器，支持 CSS/XPath 选择器、批量抓取、自动分页、数据导出（CSV/JSON）。 适用于市场调研、竞品分析、内容聚合。）
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This mostly looks like a normal web-scraping/export skill, but its config unexpectedly declares local command execution access that is not explained by the skill description.
Before installing, verify whether the platform will honor the exec capability listed in config.json. For a web data extractor, web_fetch/read/write are expected, but local command execution is not clearly justified here. If possible, remove or deny exec and use the scraper only on authorized public websites with conservative rate limits.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If this capability were honored by the runtime, the skill could give an agent local command execution access beyond what a user would expect from a web data extractor.
The config declares exec, a broad local command execution capability, while the SKILL.md frontmatter and registry requirements only list web_fetch, read, and write. The skill documentation does not explain any need to run local commands.
"capabilities": [
"web_fetch",
"read",
"write",
"exec"
]Do not grant exec for this skill unless the author documents a clear, limited need for it. Prefer removing exec from config.json or requiring explicit user approval for any command execution.
The agent may make multiple web requests and write exported data files, which can affect target websites or create unwanted local files if used carelessly.
The skill is designed to fetch multiple pages and process URL lists. This is purpose-aligned for a scraper, but users should recognize it may generate repeated requests to target websites.
### 3. 批量抓取 - 自动分页处理 - URL 列表批量处理 - 并发控制
Use only on permitted public data, limit page counts and concurrency, respect robots.txt and site terms, and review export paths before writing files.
It is harder to verify the author or project history before trusting the skill.
The package has limited provenance information. There is no code or install script in the provided artifacts, so this is a provenance note rather than evidence of malicious behavior.
Source: unknown Homepage: none
Review the provided SKILL.md and config.json before use, and prefer a version with a verifiable source repository or maintainer information.