[FEATURE]: New flag/option --read #8976

Open
opened 2026-02-16 18:11:18 -05:00 by yindo · 1 comment
Owner

Originally created by @clrblind on GitHub (Feb 10, 2026).

Originally assigned to: @thdxr on GitHub.

Feature hasn't been suggested before.

  • I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

I suggest adding the following option: --read . The purpose of this option is to mix instructions from the specified file, for example --read AGENTS.md, into each prompt within a session. This will allow the model to never lose critical conditions in a large amount of context. Of course, this will increase overhead costs, but at the same time it will allow for incredible accuracy in operation.

Originally created by @clrblind on GitHub (Feb 10, 2026). Originally assigned to: @thdxr on GitHub. ### Feature hasn't been suggested before. - [x] I have verified this feature I'm about to request hasn't been suggested before. ### Describe the enhancement you want to request I suggest adding the following option: **--read <filename>**. The purpose of this option is to mix instructions from the specified file, for example **--read AGENTS.md**, into each prompt within a session. This will allow the model to never lose critical conditions in a large amount of context. Of course, this will increase overhead costs, but at the same time it will allow for incredible accuracy in operation.
yindo added the discussion label 2026-02-16 18:11:18 -05:00
Author
Owner

@github-actions[bot] commented on GitHub (Feb 10, 2026):

This issue might be related to or overlap with existing feature requests. Please review these before implementing:

  • #4479: Configurable instruction file search boundary and file types
  • #6316: Context Auto-Discovery - Load AGENTS.md from directories when accessing files
  • #10096: Smart Rules - Context-aware instruction injection with Claude Code compatibility
  • #4479: Configurable instruction file search boundary and file types

The --read flag approach you're suggesting is a distinct implementation strategy from some of these, but coordination could help avoid duplication.

@github-actions[bot] commented on GitHub (Feb 10, 2026): This issue might be related to or overlap with existing feature requests. Please review these before implementing: - #4479: Configurable instruction file search boundary and file types - #6316: Context Auto-Discovery - Load AGENTS.md from directories when accessing files - #10096: Smart Rules - Context-aware instruction injection with Claude Code compatibility - #4479: Configurable instruction file search boundary and file types The `--read` flag approach you're suggesting is a distinct implementation strategy from some of these, but coordination could help avoid duplication.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: anomalyco/opencode#8976