skills/jpcaparas/skills/modernize-ui/Gen Agent Trust Hub

modernize-ui

Pass

Audited by Gen Agent Trust Hub on Nov 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill instructs the agent to use sudo for the command playwright install-deps <browser>. This is a standard requirement for installing necessary system libraries on Linux to allow Playwright's browser engines to function.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it retrieves content from external URLs during the UI audit process.
  • Ingestion points: External website data is loaded via page.goto(url) in scripts/capture.py.
  • Boundary markers: No explicit boundary markers or "ignore instructions" delimiters are used when the agent processes logs or measurements from these external pages.
  • Capability inventory: The skill performs network operations via Playwright, writes audit results to disk (manifest.json, screenshots), and the agent is instructed to modify the local project's code to implement approved UI changes.
  • Sanitization: scripts/build_catalogue.py uses html.escape (aliased as _e) to sanitize data ingested from the external audit (such as console errors and accessibility violations) before it is rendered into the final HTML catalogue.
  • [COMMAND_EXECUTION]: The skill relies on several CLI tools to perform its core functions.
  • Uses python and pip for running audit scripts and managing browser automation environments.
  • Uses npm and npx to execute accessibility scans with axe-core and static compatibility checks with doiuse and browserslist.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 11, 2026, 09:03 AM