coretax-automation-specialist
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses
chrome-devtools.evaluate_javascriptto execute scripts within the browser context. This is used for legitimate automation tasks such as filling form fields, triggering DOM events, and extracting transactional data from the CoreTax portal. - [EXTERNAL_DOWNLOADS]: The skill references external Model Context Protocol (MCP) tools, specifically
chrome-devtools-mcpanddesktop-rpa-computer-use, to facilitate browser interaction and operating system-level tasks. - [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it reads and processes data from external web pages.
- Ingestion points: Web content is ingested via
chrome-devtools.evaluate_javascriptand console log monitoring. - Boundary markers: The instructions do not define specific delimiters or headers to separate ingested web data from the agent's internal instructions.
- Capability inventory: The agent can execute JavaScript in the browser, capture screenshots, and write logs to local Excel and CSV files.
- Sanitization: There is no mention of sanitizing or validating data retrieved from the web portal before it is used for subsequent logic or logged to files.
Audit Metadata