create-chatgpt-mockup
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill relies on the Playwright library which requires the installation of browser binaries. The
screenshot.jsutility includes logic to trigger the download and installation of the Chromium browser using the commandnpx playwright install chromiumfrom Microsoft's official package registry if it is not present at runtime. - [COMMAND_EXECUTION]: The
screenshot.jsscript utilizesexecSyncto run the Playwright installation command. This is an expected initialization step for tools utilizing browser-based screenshot capabilities. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data from
thread.jsonand interpolates it into an HTML template for rendering. While this constitutes an indirect prompt injection surface, the risk is mitigated by built-in sanitization. - Ingestion points: The
render.jsentry point reads thethread.jsoninput file, which is then parsed and processed bygenerate.js. - Boundary markers: The skill does not explicitly define delimiters in its instructions to separate user-provided content from the agent's context during the mockup composition phase.
- Capability inventory: The skill has the ability to write files to the local disk (
index.html,screenshot.png) and uses a headless browser (Playwright) to render local and remote content. - Sanitization: The
generate.jsfile implements a robustescapeHTMLfunction that sanitizes special characters (e.g.,<,>,&,") before text is interpolated into the HTML structure, effectively preventing cross-site scripting (XSS) or HTML injection within the generated mockup.
Audit Metadata