vision-click
Warn
Audited by Snyk on May 12, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 1.00). The skill captures screenshots of whatever page is open via agbrowse/browserCmd (e.g., browserCmd(['screenshot','--json']) in vision-click.mjs and the SKILL.md pipeline) and passes those images to the Codex CLI (codex exec -i --json) for visual analysis, so arbitrary/untrusted public webpage content can influence the model's coordinate output and therefore subsequent automated mouse clicks.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata