skills/zackbart/skills/sparkle-win/Gen Agent Trust Hub

sparkle-win

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides technical documentation for integrating the legitimate and widely-used WinSparkle framework.
  • [PROMPT_INJECTION]: No malicious instruction patterns or attempts to bypass agent safety constraints were detected.
  • [EXTERNAL_DOWNLOADS]: References to external code are limited to the official project repository on GitHub and official package registries like NuGet and PyPI.
  • [REMOTE_CODE_EXECUTION]: The skill guides the implementation of an auto-update system, providing detailed security measures such as EdDSA signature verification and mandatory HTTPS to ensure the integrity of update payloads.
  • [COMMAND_EXECUTION]: Examples demonstrate the use of developer-focused command-line tools for cryptographic key generation and binary signing.
  • [DATA_EXFILTRATION]: No patterns of unauthorized data access or transmission to unknown domains were found; network operations are restricted to the updater's intended functionality.
  • [CREDENTIALS_UNSAFE]: No hardcoded secrets or private keys are present; example public keys are provided for signature verification purposes.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 11:29 PM