sparkle-win

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill includes explicit examples that embed tokens, keys, and signature strings verbatim (e.g., Authorization "Bearer my-token", EdDSA public key and sparkle:edSignature values) and instructs adding them into code/XML, which encourages the LLM to output secret values directly.

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

• Third-party content exposure detected (high risk: 1.00). The skill instructs WinSparkle to fetch and parse an arbitrary appcast feed URL (set via win_sparkle_set_appcast_url in SKILL.md and referenced in "How Update Checking Works" / "Appcast URL Configuration"), and the feed's items, releaseNotesLink, and enclosure URLs are untrusted third‑party content that directly influence update checks and can trigger downloads/auto-installs.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). Yes — WinSparkle fetches an appcast at runtime from the configured URL (e.g., https://example.com/appcast.xml) and may download/run enclosure payloads (e.g., https://example.com/setup.exe), so remotely hosted XML and installer URLs directly control the updater UI/behavior and can cause execution of remote code.