kuroco-admin-api-browser
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
javascript_toolto generate and execute JavaScript snippets within the browser context. This functionality is intended for programmatic interaction with the Kuroco Admin API using the user's active session. - [EXTERNAL_DOWNLOADS]: The documentation provides links to the official Google Chrome Web Store for the installation of the necessary
claude-in-chromeextension, which is a recognized and trusted source. - [PROMPT_INJECTION]: The skill processes content from external web pages and API responses, creating an attack surface for indirect prompt injection.
- Ingestion points: Data is ingested via
get_page_textfrom/llms.txtand through JSON responses from various Admin API endpoints. - Boundary markers: The prompt instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when interpolating external data into the agent's context.
- Capability inventory: The agent has the capability to navigate URLs, read page accessibility trees, and execute arbitrary JavaScript within the browser.
- Sanitization: The skill relies on security filtering logic within the
claude-in-chromeMCP tool to block sensitive data and provides specific instructions for the agent to filter out personal information from its responses.
Audit Metadata