Back to directory
Testing & Reviewveltiq/groundtruth
@veltiq/groundtruth
The human-in-the-loop for AI coding, automated. Catch when your agent lies, leaves typos, or skips work — verify every turn against the real diff, then make it run/screenshot/test and self-fix before finishing. Claude Code hook + CLI.
Suggested install command
npx skills add veltiq/groundtruth/@veltiq/groundtruthAlways inspect the linked repository and skill instructions before running commands. Skills are instructions; permissions and execution still matter.
Compatibility
Agent support matrix
3 supported
| Agent | Status |
|---|---|
| Claude Code | Supported |
| OpenCode | Not listed |
| Cursor | Supported |
| MCP | Not listed |
| GitHub Copilot | Not listed |
| Windsurf |