creating-bookmarklets
Pass
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill's primary purpose is generating browser-executable JavaScript (bookmarklets), which are executed in the user's browser context.
- [EXTERNAL_DOWNLOADS]: Fetches the Terser library from the jsDelivr CDN (a well-known service) to perform code minification.
- [EXTERNAL_DOWNLOADS]: Directs users to an external installation utility hosted on the author's personal domain (austegard.com).
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it transforms natural language user requests directly into executable JavaScript code.
- Ingestion points: User-provided descriptions of bookmarklet functionality in SKILL.md.
- Boundary markers: Absent; there are no instructions to delimit user input or treat it as untrusted during the code generation process.
- Capability inventory: The skill generates and provides raw JavaScript code meant for immediate browser execution.
- Sanitization: Absent; the workflow does not specify validation or sanitization of user-requested logic before delivery.
Audit Metadata