Skip to main content
Agents are best when the task needs code, tools, or per-input adaptation instead of a fixed workflow graph. The source lives in Git, so you work with normal files and commits.

​
1. Clone the agent source

Use a directory name that makes the organization and environment obvious.

​
2. Edit the agent

Open the agent files, change the prompt, tools, or code, then run any local checks your project defines.

​
3. Commit, push, and release

save validates, commits, and pushes local changes. release creates an immutable source tag, and sync points the automation at the latest release.

​
4. Run the agent

The command prints a run id. Poll or watch it like any other run:

​
5. Improve with reviews

When a run is wrong, add a review and promote the corrected case into the dataset:
The next experiment can now catch the same miss before it ships again. For a full review workflow, see Review production runs.