wp-admin-browser

Pass

Audited by Gen Agent Trust Hub on Aug 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill uses the evaluate_script tool to run JavaScript snippets directly in the browser environment. This is used for checking application state, reading localized variables, and clearing localStorage.
  • Evidence: Multiple examples in SKILL.md and references/browser-debug-patterns.md demonstrate running scripts like () => ({ myPlugin: typeof window.MY_PLUGIN, ... }) and using fetch within the browser context to obtain REST nonces.
  • [PRIVILEGE_ESCALATION]: The skill instructions require the agent to create temporary accounts with the 'Administrator' role to perform tests, which grants the agent full control over the WordPress instance.
  • Evidence: SKILL.md (Step 2) and references/wp-admin-navigation.md provide a specific workflow for creating a new user with the administrator role.
  • [INDIRECT_PROMPT_INJECTION]: The skill interacts with and reads data from the WordPress admin interface, which is a potential surface for injection if the interface contains malicious content from third-party plugins or compromised posts.
  • Ingestion points: Page content, admin notices, and DOM elements retrieved via take_snapshot and evaluate_script.
  • Boundary markers: The skill does not provide specific instructions to the agent to distinguish between UI text and system instructions.
  • Capability inventory: Access to click, fill_form, evaluate_script (with network access), and upload_file tools.
  • Sanitization: There is no mention of filtering or validating content extracted from the browser DOM before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 22, 2026, 09:07 AM