Skip to main content

runtm approve

Apply agent-proposed changes from runtm.requests.yaml to your manifest.

Arguments

Options

How It Works

AI agents can propose changes by creating a runtm.requests.yaml file:
When you run runtm approve, the CLI:
  1. Reads runtm.requests.yaml
  2. Merges changes into runtm.yaml
  3. Deletes runtm.requests.yaml

Examples

Preview Changes

Output:

Apply Changes

Output:

Requests File Format

Agent Workflow

The typical workflow:
  1. Agent creates requests - AI proposes changes via runtm.requests.yaml
  2. Human reviews - Use runtm approve --dry-run to preview
  3. Human approves - Run runtm approve to apply
  4. Human sets secrets - Use runtm secrets set for any new secrets
  5. Deploy - Run runtm deploy
This ensures humans stay in the loop while agents can propose infrastructure changes.

v1 Note

In v1, runtm approve is informational. Deploys work without approval. Future versions will enforce policy compliance for org deployments.