rever

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to fetch full request headers/bodies (which can contain API keys, bearer tokens, cookies, or passwords) and to "export_python_client" / "create_macro" to emit runnable clients — operations that require reading and embedding secret values verbatim, creating an exfiltration risk.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The GitHub repository and its releases point to a personal/unverified GitHub project that distributes a desktop app (likely binary releases) — a higher-risk distribution channel for executables; the other URLs (claude.ai/code, example.com) are benign/official or non-download pages.

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

• Third-party content exposure detected (medium risk: 0.30). At runtime the skill’s workflow (via browser_navigate, browser_snapshot, list_requests, and fetch_body) ingests free text that appears in the controlled browser tab and its captured HTTP traffic/response bodies—data authored by whoever controls the visited site/content—without requiring the agent to pre-select a specific trusted item.