extract-webpage-data
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behavior or security risks were identified. The skill correctly implements its stated purpose of web data extraction.
- [COMMAND_EXECUTION]: The skill utilizes the 'orth' command-line interface to execute scraping tasks. These executions are limited to the vendor-provided utility and are used for their intended purpose.
- [DATA_EXFILTRATION]: While the skill sends URLs to external services, this is necessary for web scraping and does not involve the unauthorized transmission of sensitive local files or credentials.
- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests untrusted data from the web for processing. This is a common characteristic of web-scraping agents.
- Ingestion points: 'url_to_scrape', 'website_url', and 'url' parameters defined in SKILL.md.
- Boundary markers: None explicitly defined in the instructions to isolate scraped content from agent instructions.
- Capability inventory: Use of the 'orth' CLI tool for network operations.
- Sanitization: No explicit sanitization of ingested web content is performed by the skill itself.
Audit Metadata