openai / openai/codex

[Linux] @ skill picker does not open although skills are installed and available on web

Open
#42,481 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

app bug Linux skills
Dominant language
Rust
Stars
125k
Forks
19.4k
PR merge metrics
PR metrics pending

Description

What version of the Codex App are you using (From “About Codex” dialog)?

26.831.20005

What subscription do you have?

ChatGPT Team

What platform is your computer?

Linux 6.8.0-138-generic x86_64 x86_64

What issue are you seeing?

Installed Skills are visible and enabled in:

Plugins -> Skills -> Installed

However, in a normal ChatGPT conversation in the Linux Desktop app,
typing @ or a Skill prefix such as @m3 does not open the Skill
picker/autocomplete.

The text remains plain text in the composer.

Opening the exact same conversation with the same account in
chatgpt.com works correctly: typing @m3 immediately opens the Skill
picker and shows the matching installed Skills.

Steps to reproduce
  1. Open ChatGPT Desktop on Linux.
  2. Open Plugins -> Skills.
  3. Verify that personal Skills are installed and enabled.
  4. Open a normal ChatGPT conversation.
  5. Type @ or a partial installed Skill name, for example @m3.
  6. Observe that no Skill picker appears.
  7. Open the exact same conversation with the same account in chatgpt.com.
  8. Type @m3.
  9. Observe that the Skill picker immediately appears and lists the
    matching installed Skills.
Expected behavior

Typing @ or @<skill-prefix> in the Linux Desktop composer should
open the Skill picker, as it does in the web client.

Selecting a Skill from the picker should attach that Skill to the
message.

Actual behavior

No Skill picker opens in the Linux Desktop app.

@m3 remains ordinary text.

The same account, same installed Skills, and same conversation work
correctly in the browser.

Reproducibility

Consistently reproducible.

Additional information

This appears to be a Desktop composer / mention-picker issue rather
than a Skill installation or synchronization issue because:

  • The Skills are visible under Plugins -> Skills -> Installed.
  • They are enabled in the Linux app.
  • The exact same conversation works correctly in the web client.
  • The browser immediately displays the expected Skill suggestions.

Related issues with Desktop mention/picker behavior:

  • #22735
  • #22716
Image Image Image
What steps can reproduce the bug?
  1. Open ChatGPT Desktop on Linux.
  2. Open Plugins -> Skills.
  3. Verify that personal Skills are installed and enabled.
  4. Open a normal ChatGPT conversation.
  5. Type @.
  6. Type part of an installed Skill name, for example @m3.
  7. Observe that no Skill picker appears and @m3 remains plain text.
  8. Open the exact same conversation with the same account in chatgpt.com.
  9. Type @m3.
  10. Observe that the Skill picker immediately appears and lists the matching installed Skills.

The issue is consistently reproducible in the Linux Desktop app.

No session ID, token limit, or context-window condition appears to be involved; the behavior can be reproduced in the composer directly.

What is the expected behavior?

Typing @ or @<skill-prefix> in the Linux ChatGPT Desktop composer should open the Skill picker, as it does in the web client.

For example, typing @m3 should show matching installed Skills such as M3 Codex Workflow, M3 Handoff, and M3 Grill Me.

Selecting an entry should attach the selected Skill to the message rather than leaving the text as an ordinary @ string.

Additional information

The issue appears to be specific to the Linux Desktop composer.

The installed Skills are visible and enabled in the app, and the same account and exact same conversation work correctly in the web client.

This makes a Skill installation or account synchronization issue unlikely.

Related Desktop mention/picker issues:

  • #22735
  • #22716

In this Linux case, however, the Skill picker does not open at all when typing @ or a Skill prefix such as @m3.

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

Start by reproducing the issue in the Linux Desktop composer with an installed Skill, then compare its @ mention and Skill-picker behavior with the web client. Trace the Desktop composer and picker entry points, using related issues #22735 and #22716 for context. Done means @ and prefixes such as @m3 show matching installed Skills and selecting one attaches it to the message.

Written by the indexing model from the issue text.

Assessment

Tech stack
linux
Domain
desktop
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
55/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.