prompts-gpt setup
Scaffold a local orchestration config file for running prompts through AI agents.
Usage
prompts-gpt setup [options]Options
| Flag | Description |
|---|---|
--prompt-file <path> | Default prompt file for run |
--prompt-files <list> | Default batch files for run-batch |
--manifest <path> | Manifest as batch source |
--prompt-dir <path> | Scan directory for prompts |
--agent <name> | Default orchestration profile. Default: router |
--provider-order <list> | Router provider priority |
--timeout <seconds> | Default timeout. Default: 900 |
--retry-count <n> | Retry count. Default: 0 |
--artifacts-dir <path> | Artifacts directory. Default: .scripts/runs |
--codex-model <name> | Model for Codex |
--cursor-model <name> | Model for Cursor |
--claude-model <name> | Model for Claude |
--copilot-model <name> | Model for Copilot |
--allow-destructive-git | Disable destructive git protection |
--overwrite | Replace existing config |
--json | Output config as JSON |
What It Creates
.prompts-gpt/config.json:
{
"defaultAgent": "router",
"providerOrder": ["codex", "cursor", "claude", "copilot"],
"timeoutSeconds": 900,
"retryCount": 0,
"artifactsDir": ".scripts/runs",
"modelOverrides": {},
"disallowDestructiveGit": true
}