agent-browser

Fail

Audited by Snyk on Mar 7, 2026

Risk Level: HIGH
Full Analysis

HIGH W007: Insecure credential handling detected in skill instructions.

  • Insecure credential handling detected (high risk: 1.00). The prompt includes examples and workflows that embed credentials directly into CLI commands (e.g., agent-browser fill @e2 "password123" and fill with "$PASSWORD"), which require the agent to output secret values verbatim and therefore pose an exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 1.00). The skill explicitly navigates to arbitrary external URLs (SKILL.md "Navigate: agent-browser open ") and the templates and command references (e.g., templates/capture-workflow.sh, snapshot -i, get text body, eval) ingest and act on page content, so untrusted public webpages can directly influence agent decisions and actions.
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 7, 2026, 03:42 PM