budge
Fail
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires downloading a remote JavaScript file from https://www.budge.design/budge.iife.js. This domain is not recognized as a trusted source.
- [REMOTE_CODE_EXECUTION]: The skill instructs the agent to modify the project's root layout (app/layout.tsx) to include a remote script. This script executes in the end-user's browser, providing a vector for arbitrary code execution in the client context.
- [COMMAND_EXECUTION]: The agent is directed to use StrReplace to programmatically modify the application's layout and component files, adding both the remote script and configuration elements.
- [PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by instructing the agent to follow instructions generated by the remote widget and pasted by the user.
- Ingestion points: The user pastes text into the agent's chat window that was copied to the clipboard by the external script.
- Boundary markers: No markers or warnings are used to distinguish the generated text from legitimate user instructions.
- Capability inventory: The agent has the capability to perform file system modifications via StrReplace.
- Sanitization: There is no validation or sanitization of the content provided in the "submitted prompt" before the agent acts upon it.
Recommendations
- AI detected serious security threats
Audit Metadata