Why is Mode specific to triggers ?

[K3UL]

I don't understand why the mode is restricted to the trigger type.
For example, why can't I use Agent Mode on a PR to do some automations silently ?

Moreover, you mention in your v1.0 vision that the mode should be auto-detected, and I think this should be an option (mode could have the values agent, tag, review, auto and would default to auto), because otherwise it will just goes further to break/prevent advanced automations.

[Advait251206]

From my understanding, the current trigger-to-mode mapping is primarily a product design decision rather than a technical limitation.

Each trigger has a well-defined purpose:

• Review mode is intended for pull request reviews, where Claude provides feedback through GitHub's review system.
• Agent mode is designed for interactive tasks where Claude can make changes, open commits, or work through a requested task.
• Tag mode is optimized for workflows that are explicitly invoked by mentioning or tagging Claude.

By coupling the mode to the trigger, the action can provide predictable behavior and reduce configuration complexity. It also helps avoid ambiguous cases where the same workflow could behave differently depending on the selected mode.

That said, I think your use case is a valid one. Running Agent mode automatically on PR events (without requiring a mention) could enable useful automation scenarios such as:

• Automatically fixing formatting or lint issues.
• Updating generated files.
• Applying straightforward refactorings.
• Keeping documentation in sync.
• Performing repetitive maintenance tasks.

Having an explicit mode input (for example: agent, review, tag, or auto) that defaults to auto would preserve the current behavior while allowing advanced users to override it when they understand the implications. That would also align well with the long-term vision of automatic mode detection while still providing an escape hatch for more sophisticated workflows.

So I don't think there's an inherent technical reason why Agent mode couldn't be used with PR triggers—it appears to be an intentional design choice in the current implementation. Whether that flexibility should be exposed is ultimately up to the maintainers, but I agree it would make the Action more powerful for advanced automation.

---

If this answer helped or pointed you in the right direction, I'd appreciate it if you could mark it as the accepted answer so it's easier for others with the same issue to find.

Also, if you found my contribution useful, I'd appreciate it if you could check out my GitHub profile, follow me, and star any repositories you find interesting.

GitHub: https://github.com/Advait251206