skills/unbrowse-ai/sites/kompas/Gen Agent Trust Hub

kompas

Pass

Audited by Gen Agent Trust Hub on Sep 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill connects to and installs tools from the vendor's official service endpoint at unbrowse.ai (README.md, SKILL.md).
  • [DYNAMIC_EXECUTION]: The skill utilizes a remote service that can dynamically compile new tools based on agent actions in a cloud browser session, which are subsequently listed for the agent's use (SKILL.md).
  • [INDIRECT_PROMPT_INJECTION]: The skill processes live news content and search results, presenting a surface for potential adversarial instructions.
  • Ingestion points: Retrieves news articles and search result listings from kompas.com via tools such as kompas_com__get_search and kompas_com__get_list (SKILL.md).
  • Boundary markers: No explicit delimitation or "ignore" instructions for embedded data are specified in the prompt instructions.
  • Capability inventory: The agent has the capability to search news, read articles, and browse category listings, as well as perform broader site-related tasks via unbrowse.run (SKILL.md).
  • Sanitization: There is no documented evidence of content sanitization or validation for the text retrieved from the external news site.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 28, 2026, 10:56 AM