speckit-updater
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill's instructions include hardcoded absolute paths to a specific user directory (
C:\Users\bobby\.claude\skills\speckit-updater\scripts\update-wrapper.ps1) for script execution. This is a non-standard practice that creates environment-specific dependencies and potential execution failures for other users. - [EXTERNAL_DOWNLOADS]: The skill fetches version metadata and project templates from the GitHub Releases API and external repositories at runtime.
- [REMOTE_CODE_EXECUTION]: The skill is designed to self-update by downloading and replacing its own PowerShell scripts from GitHub. While it uses a well-known service (GitHub), the pattern of downloading and immediately executing modified scripts is a known risk vector for code execution.
- [PROMPT_INJECTION]: The skill processes untrusted data from the GitHub API and local project files (e.g., manifest and templates) to drive its internal logic. This ingestion point (GitHub Releases API and project manifest) lacks explicit sanitization or boundary markers, and combined with capabilities like PowerShell execution and file writes, it creates a surface for indirect prompt injection if external repository data is compromised.
Audit Metadata