agent-browser

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt includes explicit examples and workflows that embed plaintext credentials into CLI commands and fill actions (e.g., agent-browser fill @e2 "password123", echo "pass" | agent-browser auth save ...), which would require an LLM-driven agent to output secret values verbatim, creating an exfiltration risk.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). Although many listed URLs are benign placeholders (example.com, github.com/login), local host, or documentation pages (lightpanda.io), the set includes explicit "malicious.com" and several unknown domains (site-a/site-b, prod/staging/app subdomains) that could host or redirect to untrusted installers — and the skill instructs downloading/executing from these locations — so the overall package is moderately to highly suspicious without further provenance or checks.

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

• Third-party content exposure detected (high risk: 1.00). The skill's core workflow (SKILL.md) and templates (e.g., templates/capture-workflow.sh and templates/form-automation.sh) explicitly instruct the agent to "open ", take snapshots, and "get text body"/"snapshot -i" from arbitrary web pages, meaning it fetches and reads untrusted public third‑party content that can directly influence subsequent actions.