test-error-states
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to install and run the
@reticlehq/serverpackage from the vendor's repository usingnpx. This is a vendor-owned tool used for initializing the execution environment. - [COMMAND_EXECUTION]: The documentation includes a command to launch Google Chrome with remote debugging enabled (
--remote-debugging-port=9222). This is a standard procedure for allowing automation tools to interface with a local browser instance. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to interact with and process data from external web applications, which introduces a surface for indirect prompt injection.
- Ingestion points: The skill ingests data from the browser's Document Object Model (DOM) and the developer console (e.g., checking for
error-bannerelements or specificconsolelog levels) during automated interactions. - Boundary markers: The instructions do not define explicit delimiters to separate system instructions from the untrusted data being retrieved from the browser.
- Capability inventory: The skill has access to tools for browser automation, network request interception (
reticle_network_mock), and clock manipulation (reticle_clock). - Sanitization: There is no evidence of specific sanitization or filtering logic applied to the content retrieved from the browser before it is processed by the agent.
Audit Metadata