agent-browser

Fail

Audited by Snyk on Feb 18, 2026

Risk Level: HIGH
Full Analysis

HIGH W007: Insecure credential handling detected in skill instructions.

  • Insecure credential handling detected (high risk: 0.90). The prompt includes examples and commands that place passwords, header values, cookies, and credentials directly into CLI arguments (e.g., fill @e2 "password123", set credentials user pass, set headers '{"X-Key":"v"}'), which would require the agent to emit secret values verbatim and thus poses 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.md shows the agent explicitly navigates to arbitrary URLs (agent-browser open ), snapshots and reads page content (agent-browser snapshot -i, get text/get html) and even evaluates page JS (eval), so it fetches and interprets untrusted public web content which can directly drive subsequent actions.
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 18, 2026, 08:47 PM