You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[Question]: why not support symbol/function @mentions in prompt ? #11878
Hi all — maybe I missed something, but I can’t find a quick way to @mention a function/method/class symbol directly in OpenCode.
In Cursor/Windsurf I can just @ a symbol and it’s immediately part of the context. In OpenCode, I always end up going back to Neovim to search the symbol and manually copy/paste it, which is pretty inconvenient.
This is a large open-source project, so I doubt I’m the first person asking for this. I’m not asking for implementation here — just raising the need: symbol-level @mentions, not just file/agent.
I also had GPT investigate this requirement, and it seems that current custom tools/plugins don't yet support custom mention sources. However, I haven't looked into this issue in depth yet; if it does, I'll probably implement one for myself.
If this isn’t supported for a reason (limitations or tradeoffs), I’d love to know.
Reacted by imadNaseer, Neyas Guruswamy, Jose Albizures, Nikolai Krapyvchenko and Jeremy Morrison
When dealing with behemoths like MySQL, 99% of your effort goes into tracing complex call chains and understanding module interactions. Conversations with the agent become an iterative cycle of exploration and reference. To maintain the agent's reasoning foundation and prevent it from veering off course, you must frequently provide it with specific symbols and code snippets.
The bottleneck here is context-switching overhead. Every time I have to pause my train of thought and manually jump between buffers, copy and paste symbols into the prompt, it creates friction.
Activity
github-actions commentedon Feb 3, 2026
This issue might be a duplicate of existing issues. Please check:
Feel free to ignore if none of these address your specific case.
jensenojs commentedon Feb 3, 2026
When dealing with behemoths like MySQL, 99% of your effort goes into tracing complex call chains and understanding module interactions. Conversations with the agent become an iterative cycle of exploration and reference. To maintain the agent's reasoning foundation and prevent it from veering off course, you must frequently provide it with specific symbols and code snippets.
The bottleneck here is context-switching overhead. Every time I have to pause my train of thought and manually jump between buffers, copy and paste symbols into the prompt, it creates friction.
update :
imadNaseer commentedon Feb 18, 2026
got here searching for the exact same feature +1
github-actions commentedon Apr 20, 2026
To stay organized issues are automatically closed after 60 days of no activity. If the issue is still relevant please open a new one.
jensenojs commentedon May 16, 2026
reopen it