scraperapi-java-sdk

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates indirect prompt injection by design, as it provides methods to ingest untrusted data from external websites.
  • Ingestion point: The client.get(url).result() method in SKILL.md fetches arbitrary content from user-provided URLs.
  • Boundary markers: The provided code examples do not demonstrate the use of delimiters or 'ignore' instructions to isolate scraped content from the agent's logic.
  • Capability inventory: The skill utilizes ScraperApiClient and java.net.http.HttpClient for network operations in SKILL.md.
  • Sanitization: Examples return raw HTML strings or JSON without demonstrating sanitization or filtering of the external data.
  • [COMMAND_EXECUTION]: The skill provides instructions for executing Java code that performs network requests to ScraperAPI's official infrastructure (scraperapi.com, async.scraperapi.com) to facilitate web scraping.
  • [EXTERNAL_DOWNLOADS]: The documentation references the com.scraperapi:sdk Maven/Gradle artifact and the Jackson JSON library for project setup and dependency management.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 08:27 PM
Security Audit — agent-trust-hub — scraperapi-java-sdk