openai / openai/codex

[IDE Extension] Repeated unnecessary reads of MEMORY.md during an active conversation

Open
#41,621 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug extension memory
Dominant language
Rust
Stars
125k
Forks
19.4k
PR merge metrics
PR metrics pending

Description

What version of the IDE extension are you using?

26.825.51511

What subscription do you have?

chatgpt pro

Which IDE are you using?

VS Code

What platform is your computer?

macos

What issue are you seeing?

Within a single continuous conversation, Codex repeatedly runs filesystem searches against ~/.codex/memories/MEMORY.md before answering follow-up questions, even when all relevant context is already present in the current thread.

These internal memory reads are visible as tool calls in the IDE extension. They add unnecessary latency and activity noise, and make it appear that Codex is performing unrelated filesystem operations.

MEMORY.md was only read, not modified.

What steps can reproduce the bug?
  1. Open a Codex conversation in the IDE extension.
  2. Discuss a repository or implementation over several consecutive turns.
  3. Ask a simple follow-up question that relies entirely on the current conversation.
  4. Observe Codex searching or reading ~/.codex/memories/MEMORY.md before answering.
  5. Repeat with additional follow-up questions; the memory lookup may occur again.
What is the expected behavior?

Codex should use the current conversation context when it is sufficient.

Persistent memory should only be consulted when the current context is missing necessary information, or when the user explicitly asks to recall information from another session. Internal memory operations should not repeatedly appear as unrelated tool activity in the IDE extension.

Additional information

This occurs within one uninterrupted conversation, not after resuming or switching sessions. The complete relevant context remains visible in the current thread.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the repeated ~/.codex/memories/MEMORY.md reads in the VS Code IDE extension during consecutive follow-up questions. Trace the memory lookup entry point and compare behavior when the current conversation already contains the needed context. Done means persistent memory is consulted only when context is missing or explicitly requested, without repeated unrelated tool activity.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust, vscode
Domain
developer-experience, tooling
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.