Forums

Articles
Create
cancel
Showing results for 
Search instead for 
Did you mean: 

Best practices for storing and versioning ROVO Agent instructions in GitHub?

Sameer Sawant
June 8, 2026

Hi Atlassian Community,

Our team has been exploring ways to manage ROVO Agent instructions more systematically, and we're wondering if anyone has experience storing and versioning them in a GitHub repository.

Specifically, we'd love to know:

  1. Is there an official way to export ROVO Agent configurations outside of the Atlassian UI?
  2. Has anyone set up a GitHub-based workflow to version-control their ROVO Agent instructions?
  3. Are there any Atlassian APIs that allow pushing or syncing agent instructions programmatically?
  4. What's your team's current approach to tracking changes to ROVO Agent instructions over time?

Our goal is to treat agent instructions as "configuration as code" — with proper versioning, peer review via PRs, and an audit trail.

Any community experiences, workarounds, or official guidance would be greatly appreciated!

Thanks in advance 🙏

1 answer

0 votes
Alexander Nilsson
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Champions.
June 8, 2026

Hello Sameer,

thanks for your great questions. The answer depens whether the agent was built in the Rovo Studio UI or as a Forge app, because only one of those two paths supports configuration as code today.

If you build the agent as a Forge app using the rovo:agent module, you already get the workflow you described. The agent definition lives in manifest.yml, and the instructions do not have to be inlined. The prompt property accepts either a string or a relative path to a declared resource, for example prompt: resource:agent-resource;prompts/agent-prompt.txt. That makes the prompt a plain text file you commit to GitHub, review through PRs, and ship with forge deploy. Git history becomes your audit trail, and the deploy pipeline is effectively the "push" mechanism. The full manifest reference is here:

https://developer.atlassian.com/platform/forge/manifest-reference/modules/rovo-agent/

For agents created in the Rovo Studio UI, the honest answer to your questions 1 and 3 is that there is no supported public API and no import/export today. Studio does not surface agent configurations as files, so you cannot export them, version them outside the product, or sync them programmatically. That exact gap is tracked in Atlassian's public issue tracker as ROVO-516 ("Public APIs for Rovo Agent Management"). If managing Studio agents as code matters to your team, voting and watching it is the most useful action right now:

https://jira.atlassian.com/browse/ROVO-516

So the practical recommendation: if configuration as code is a hard requirement, author those agents as Forge apps from the start. You keep the manifest and the prompt file in your repo and get PR review plus versioning for free. Studio is the faster route for non-developers, but it stays inside the Atlassian UI until that API request lands.

Looking forward to your message. 

Greetings,

Alex

Suggest an answer

Log in or Sign up to answer
DEPLOYMENT TYPE
CLOUD
PRODUCT PLAN
STANDARD
TAGS
AUG Leaders

Atlassian Community Events