Solves #700
## State of the World Before
Prior to this PR, when users wanted to share file contents with Codex,
they had two options:
- Manually copy and paste file contents into the chat
- Wait for the assistant to use the shell tool to view the file
The second approach required the assistant to:
1. Recognize the need to view a file
2. Execute a shell tool call
3. Wait for the tool call to complete
4. Process the file contents
This consumed extra tokens and reduced user control over which files
were shared with the model.
## State of the World After
With this PR, users can now:
- Reference files directly in their chat input using the `@path` syntax
- Have file contents automatically expanded into XML blocks before being
sent to the LLM
For example, users can type `@src/utils/config.js` in their message, and
the file contents will be included in context. Within the terminal chat
history, these file blocks will be collapsed back to `@path` format in
the UI for clean presentation.
Tag File suggestions:
<img width="857" alt="file-suggestions"
src="https://github.com/user-attachments/assets/397669dc-ad83-492d-b5f0-164fab2ff4ba"
/>
Tagging files in action:
<img width="858" alt="tagging-files"
src="https://github.com/user-attachments/assets/0de9d559-7b7f-4916-aeff-87ae9b16550a"
/>
Demo video of file tagging:
[](https://www.youtube.com/watch?v=vL4LqtBnqt8)
## Implementation Details
This PR consists of 2 main components:
1. **File Tag Utilities**:
- New `file-tag-utils.ts` utility module that handles both expansion and
collapsing of file tags
- `expandFileTags()` identifies `@path` tokens and replaces them with
XML blocks containing file contents
- `collapseXmlBlocks()` reverses the process, converting XML blocks back
to `@path` format for UI display
- Tokens are only expanded if they point to valid files (directories are
ignored)
- Expansion happens just before sending input to the model
2. **Terminal Chat Integration**:
- Leveraged the existing file system completion system for tabbing to
support the `@path` syntax
- Added `updateFsSuggestions` helper to manage filesystem suggestions
- Added `replaceFileSystemSuggestion` to replace input with filesystem
suggestions
- Applied `collapseXmlBlocks` in the chat response rendering so that
tagged files are shown as simple `@path` tags
The PR also includes test coverage for both the UI and the file tag
utilities.
## Next Steps
Some ideas I'd like to implement if this feature gets merged:
- Line selection: `@path[50:80]` to grab specific sections of files
- Method selection: `@path#methodName` to grab just one function/class
- Visual improvements: highlight file tags in the UI to make them more
noticeable
\+ cleanup below input help to be "ctrl+c to exit | "/" to see commands
| enter to send" now that we have command autocompletion
\+ minor other drive-by code cleanups
---------
Signed-off-by: Thibault Sottiaux <tibo@openai.com>
fix: pass correct selected model in ModelOverlay
The ModelOverlay component was incorrectly passing the current model
instead of the newly selected model to its onSelect callback. This
prevented model changes from being applied properly.
The fix ensures that when a user selects a new model, the parent
component receives the correct newly selected model value, allowing
model changes to work as intended.
## Description
This PR addresses the following improvements:
**Unify Prettier Version**: Currently, the Prettier version used in
`/package.json` and `/codex-cli/package.json` are different. In this PR,
we're updating both to use Prettier v3.
- Prettier v3 introduces improved support for JavaScript and TypeScript.
(e.g. the formatting scenario shown in the image below. This is more
aligned with the TypeScript indentation standard).
<img width="1126" alt="image"
src="https://github.com/user-attachments/assets/6e237eb8-4553-4574-b336-ed9561c55370"
/>
**Add Prettier Auto-Formatting in lint-staged**: We've added a step to
automatically run prettier --write on JavaScript and TypeScript files as
part of the lint-staged process, before the ESLint checks.
- This will help ensure that all committed code is properly formatted
according to the project's Prettier configuration.
Solves #510
This PR changes the `/bug` command to print the URL into the terminal
(so it works in headless sessions) instead of trying to open a browser.
---------
Co-authored-by: Thibault Sottiaux <tibo@openai.com>
Up-to-date of #78Fixes#32
addressed requested changes @tibo-openai :) made sense to me
though, previous rationale with passing the state up was assuming there
could be a future need to have a shared state with all available models
being available to the parent
## Description
The `as AppConfig` type assertion in the constructor may introduce
potential type safety risks. Removing the assertion and making `notify`
an optional parameter could enhance type robustness and prevent
unexpected runtime errors.
close: #605
### What
- Add support for loading and merging custom provider configurations
from a local `providers.json` file.
- Allow users to override or extend default providers with their own
settings.
### Why
This change enables users to flexibly customize and extend provider
endpoints and API keys without modifying the codebase, making the CLI
more adaptable for various LLM backends and enterprise use cases.
### How
- Introduced `loadProvidersFromFile` and `getMergedProviders` in config
logic.
- Added/updated related tests in [tests/config.test.tsx]
### Checklist
- [x] Lint passes for changed files
- [x] Tests pass for all files
- [x] Documentation/comments updated as needed
---------
Co-authored-by: Thibault Sottiaux <tibo@openai.com>
A recent commit introduced the ability to use third-party model
providers. (Really appreciate it!)
However, the usage is inconsistent: some pieces of code use the custom
providers, whereas others still have the old behavior. Additionally,
`OPENAI_BASE_URL` is now being disregarded when it shouldn't be.
This PR normalizes the usage to `getApiKey` and `getBaseUrl`, and
enables the use of `OPENAI_BASE_URL` if present.
---------
Co-authored-by: Gabriel Bianconi <GabrielBianconi@users.noreply.github.com>
## What does this PR do?
* Implements the full `/clear` command in **codex‑cli**:
* Resets chat history **and** wipes the terminal screen.
* Shows a single system message: `Context cleared`.
* Adds comprehensive unit tests for the new behaviour.
## Why is it needed?
* Fixes user‑reported bugs:
* **#395**
* **#405**
## How is it implemented?
* **Code** – Adds `process.stdout.write('\x1b[3J\x1b[H\x1b[2J')` in
`terminal.tsx`. Removed reference to `prev` in `
setItems((prev) => [
...prev,
` in `terminal-chat-new-input.tsx` & `terminal-chat-input.tsx`.
## CI / QA
All commands pass locally:
```bash
pnpm test # green
pnpm run lint # green
pnpm run typecheck # zero TS errors
```
## Results
https://github.com/user-attachments/assets/11dcf05c-e054-495a-8ecb-ac6ef21a9da4
---------
Co-authored-by: Thibault Sottiaux <tibo@openai.com>
This PR tidies up primitives under storage/.
**Noop changes:**
* Promote logger implementation to top-level utility outside of agent/
* Use logger within storage primitives
* Cleanup doc strings and comments
**Functional changes:**
* Increase command history size to 10_000
* Remove unnecessary debounce implementation and ensure a session ID is
created only once per agent loop
---------
Signed-off-by: Thibault Sottiaux <tibo@openai.com>
## What
This PR targets #402 and refactors the `history-overlay.tsx`component to
reduce cognitive complexity by splitting the `buildLists` function into
smaller, focused helper functions. It also adds comprehensive test
coverage to ensure the functionality remains intact.
## Why
The original `buildLists` function had high cognitive complexity due to
multiple nested conditionals, complex string manipulation, and mixed
responsibilities. This refactor makes the code more maintainable and
easier to understand while preserving all existing functionality.
## How
- Split `buildLists` into focused helper functions
- Added comprehensive test coverage for all functionality
- Maintained existing behavior and keyboard interactions
- Improved code organization and readability
## Testing
All tests pass, including:
- Command mode functionality
- File mode functionality
- Keyboard interactions
- Error handling
As described in
https://github.com/openai/codex/issues/392#issuecomment-2817090022
introduced by #400
The testing I'd done worked correctly because I was using the (s)
shortcut, but selecting the same option using arrow‑key → Enter on
“Switch approval mode” was preventing the user from subsequently exiting
the Switch approval mode dialog, requiring a ^C to quit codex entirely.
With this fix, both entry methods work correctly in my testing.
Per codex:
Issue
- When you navigated down (↓) to “Switch approval mode (s)” in the Shell
Command review dialog and pressed Enter, the ApprovalModeOverlay would
open—but because the underlying `TerminalChatCommandReview` component
stayed mounted (albeit disabled), its own Ink input handlers immediately
re‑captured the same key events and re‑opened the overlay as soon as you
hit Esc or Enter again. In practice this made it impossible to exit the
submenu.
Root cause
- We only disabled the SelectInput via `isDisabled`, but never fully
unmounted the review UI when an overlay was shown, so its `useInput` and
`<Select>` hooks were still active and “stealing” keys.
Fix
- In `terminal-chat.tsx` we now only render `<TerminalChatInput>` (and
by extension `TerminalChatCommandReview`) when `overlayMode === "none"`.
That unmounts all of its key handlers whenever any overlay (history,
model, approval, help, diff) is open, so no input leaks through.
Files changed
- **src/components/chat/terminal-chat.tsx**: Wrapped the entire
`<TerminalChatInput>` block in `overlayMode === "none" && agent`
With that in place, arrow‑key → Enter on “Switch approval mode”
correctly opens the overlay, and then you can use Enter/Esc inside the
overlay without getting stuck or immediately re‑opening it.
It appears that use of `isLoggingEnabled()` was erroneously copypasta'd
in many places. This PR updates its docstring to clarify that should
only be used to avoid constructing a potentially expensive docstring.
With this change, the only function that merits/uses this check is
`execCommand()`.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/420).
* #423
* __->__ #420
* #419
## Description
This PR implements multi-line input support for Codex when it asks for
user feedback (Issue #344). Users can now use Shift+Enter to add new
lines in their responses, making it easier to provide formatted code
snippets, lists, or other structured content.
## Changes
- Replace the single-line TextInput component with the
MultilineTextEditor component in terminal-chat-input.tsx
- Add support for Shift+Enter to create new lines
- Update key handling logic to properly handle history navigation in a
multi-line context
- Add reference to the editor to access cursor position information
- Update help text to inform users about the Shift+Enter functionality
- Add tests for the new functionality
## Testing
- Added new test file (terminal-chat-input-multiline.test.tsx) to test
the multi-line input functionality
- All existing tests continue to pass
- Manually tested the feature to ensure it works as expected
## Fixes
Closes#344
## Screenshots
N/A
## Additional Notes
This implementation maintains backward compatibility while adding the
requested multi-line input functionality. The UI remains clean and
intuitive, with a simple hint about using Shift+Enter for new lines.
---------
Co-authored-by: Thibault Sottiaux <tibo@openai.com>
https://github.com/openai/codex/pull/160 introduced a call to `exec()`
that takes a format string as an argument, but it is not clear that the
expansions within the format string are escaped safely. As written, it
is possible a carefully crafted command (e.g., if `cwd` were `"; && rm
-rf` or something...) could run arbitrary code.
Moving to `spawn()` makes this a bit better, as now at least `spawn()`
itself won't run an arbitrary process, though I suppose `osascript`
itself still could if the value passed to `-e` were abused. I'm not
clear on the escaping rules for AppleScript to ensure that `safePreview`
and `cwd` are injected safely.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/416).
* #423
* #420
* #419
* __->__ #416
This check was lost in https://github.com/openai/codex/pull/287. Both
the root folder and `codex-cli/` have their own `pnpm format` commands
that check the formatting of different things.
Also ran `pnpm format:fix` to fix the formatting violations that got in
while this was disabled in CI.
---
[//]: # (BEGIN SAPLING FOOTER)
Stack created with [Sapling](https://sapling-scm.com). Best reviewed
with [ReviewStack](https://reviewstack.dev/openai/codex/pull/417).
* #420
* #419
* #416
* __->__ #417
Add interactive slash‑command autocomplete & navigation in chat input
Description
This PR enhances the chat input component by adding first‑class support
for slash commands (/help, /clear, /compact, etc.)
with:
* **Live filtering:** As soon as the user types leading `/`, a list of
matching commands is shown below the prompt.
* **Arrow‑key navigation:** Up/Down arrows cycle through suggestions.
* **Enter to autocomplete:** Pressing Enter on a partial command will
fill it (without submitting) so you can add
arguments or simply press Enter again to execute.
* **Type‑safe registry:** A new `slash‑commands.ts` file declares all
supported commands in one place, along with
TypeScript types to prevent drift.
* **Validation:** Only registered commands will ever autocomplete or be
suggested; unknown single‑word slash inputs still
show an “Invalid command” system message.
* **Automated tests:**
* Unit tests for the command registry and prefix filtering
* Existing tests continue passing with no regressions
Motivation
Slash commands provide a quick, discoverable way to control the agent
(clearing history, compacting context, opening overlays,
etc.). Before, users had to memorize the exact command or rely on the
generic /help list—autocomplete makes them far more
accessible and reduces typos.
Changes
* `src/utils/slash‑commands.ts` – defines `SlashCommand` and exports a
flat list of supported commands + descriptions
* `terminal‑chat‑input.tsx`
* Import and type the command registry
* Render filtered suggestions under the prompt when input starts with
`/`
* Hook into `useInput` to handle Up/Down and Enter for selection & fill
* Flag to swallow the first Enter (autocomplete) and only submit on the
next
* Updated tests in `tests/slash‑commands.test.ts` to cover registry
contents and filtering logic
* Removed old JS version and fixed stray `@ts‑expect‑error`
How to test locally
1. Type `/` in the prompt—you should see matching commands.
2. Use arrows to move the highlight, press Enter to fill, then Enter
again to execute.
3. Run the full test suite (`npm test`) to verify no regressions.
Notes
* Future work could include fuzzy matching, paging long lists, or more
visual styling.
* This change is purely additive and does not affect non‑slash inputs or
existing slash handlers.
---------
Co-authored-by: Fouad Matin <169186268+fouad-openai@users.noreply.github.com>
Co-authored-by: Thibault Sottiaux <tibo@openai.com>
Implements https://github.com/openai/codex/issues/392
When the user is in suggest or auto-edit mode and gets an approval
request, they now have an option in the `Shell Command` dialog to:
`Switch approval mode (v)`
That option brings up the standard `Switch approval mode` dialog,
allowing the user to switch into the desired mode, then drops them back
to the `Shell Command` dialog's `Allow command?` prompt, allowing them
to approve the current command and let the agent continue doing the rest
of what it was doing without interruption.
```
╭────────────────────────────────────────────────────────
│Shell Command
│
│$ apply_patch << 'PATCH'
│*** Begin Patch
│*** Update File: foo.txt
│@@ -1 +1 @@
│-foo
│+bar
│*** End Patch
│PATCH
│
│
│Allow command?
│
│ Yes (y)
│ Explain this command (x)
│ Edit or give feedback (e)
│ Switch approval mode (v)
│ No, and keep going (n)
│ No, and stop for now (esc)
╰────────────────────────────────────────────────────────╭────────────────────────────────────────────────────────
│ Switch approval mode
│ Current mode: suggest
│
│
│
│ ❯ suggest
│ auto-edit
│ full-auto
│ type to search · enter to confirm · esc to cancel
╰────────────────────────────────────────────────────────
```
Adding in an option to turn on flex processing mode to reduce costs when
running the agent.
Bumped the openai typescript version to add the new feature.
---------
Co-authored-by: Thibault Sottiaux <tibo@openai.com>
Fix: Shift + Enter no longer prints “[27;2;13~” in the single‑line
input. Validated as working and necessary in Ghostty on Linux.
## Key points
- src/components/vendor/ink-text-input.tsx
- Added early handler that recognises the two modifyOtherKeys
escape‑sequences
- [13;<mod>u (mode 2 / CSI‑u)
- [27;<mod>;13~ (mode 1 / legacy CSI‑~)
- If Ctrl is held (hasCtrl flag) → call onSubmit() (same as plain
Enter).
- Otherwise → insert a real newline at the caret (same as Option+Enter).
- Prevents the raw sequence from being inserted into the buffer.
- src/components/chat/multiline-editor.tsx
- Replaced non‑breaking spaces with normal spaces to satisfy eslint
no‑irregular‑whitespace rule (no behaviour change).
All unit tests (114) and ESLint now pass:
npm test ✔️
npm run lint ✔️
### What is added?
I extend the if-else blocks with an additional condition where the
commands validity is checked. This only applies for entered inputs that
start with '/' and are a single word. This isn't necessarily restrictive
from the previous behavior of the program. When an invalid command is
detected, an error message is printed with a direction to retrieve
command list.
### Why is it added?
There are three main reasons for this change
**1. Model Hallucination**: When invalid commands are passed as prompts
to models, models hallucinate behavior. Since there was a fall through
in invalid commands, the models took these as prompts and hallucinated
that they completed the prompted task. An example of this behavior is
below. In the case of this example, the model though they had access to
`/clearhistory` tool where in reality that isn't the case.
A before and after effect of this tool is below:
**2. Saves Users Credits and Time**: Since false commands and invalid
commands aren't processed by the model, the user doesn't spend money on
stuff that could have been mitigated much easily. The time argument is
especially applicable for reasoning models.
**3. Saves GPU Time**: GPU time is valuable, and it is not necessary to
spend it on unnecessary/invalid requests.
### Code Explanation
If no command is matched, we check if the inputValue start with `/`
which indicated the input is a command (I will address the case where it
is isn't below). If the inputValue start with `/` we enter the else if
statement. I used a nested if statement for readability and further
extendability in the future.
There are alternative ways to check besides regex, but regex is a short
code and looks clean.
**Check Conditions**: The reason why I only check single word(command)
case is that to allow prompts where the user might decide to start with
`/` and aren't commands. The nested if statements also come in handy
where in the future other contributors might want to extend this
checking.
The code passes type, lint and test checks.
Added the ability to compact. Not sure if I should switch the model over
to gpt-4.1 for longer context or if keeping the current model is fine.
Also I'm not sure if setting the compacted to system is best practice,
would love feedback 😄
Mentioned in this issue: https://github.com/openai/codex/issues/230
Updated the position of the cursor on the user input box to be at the
end of the text when the user uses the arrow keys to navigate through
the input history in order to better match the behavior of a terminal.
This adds support for a new flag, `-w,--writable-root`, that can be
specified multiple times to _amend_ the list of folders that should be
configured as "writable roots" by the sandbox used in `full-auto` mode.
Values that are passed as relative paths will be resolved to absolute
paths.
Incidentally, this required updating a number of the `agent*.test.ts`
files: it feels like some of the setup logic across those tests could be
consolidated.
In my testing, it seems that this might be slightly out of distribution
for the model, as I had to explicitly tell it to run `apply_patch` and
that it had the permissions to write those files (initially, it just
showed me a diff and told me to apply it myself). Nevertheless, I think
this is a good starting point.
# Shell Command Explanation Option
## Description
This PR adds an option to explain shell commands when the user is
prompted to approve them (Fixes#110). When reviewing a shell command,
users can now select "Explain this command" to get a detailed
explanation of what the command does before deciding whether to approve
or reject it.
## Changes
- Added a new "EXPLAIN" option to the `ReviewDecision` enum
- Updated the command review UI to include an "Explain this command (x)"
option
- Implemented the logic to send the command to the LLM for explanation
using the same model as the agent
- Added a display for the explanation in the command review UI
- Updated all relevant components to pass the explanation through the
component tree
## Benefits
- Improves user understanding of shell commands before approving them
- Reduces the risk of approving potentially harmful commands
- Enhances the educational aspect of the tool, helping users learn about
shell commands
- Maintains the same workflow with minimal UI changes
## Testing
- Manually tested the explanation feature with various shell commands
- Verified that the explanation is displayed correctly in the UI
- Confirmed that the user can still approve or reject the command after
viewing the explanation
## Screenshots
## Additional Notes
The explanation is generated using the same model as the agent, ensuring
consistency in the quality and style of explanations.
---------
Signed-off-by: crazywolf132 <crazywolf132@gmail.com>
This PR adds a command history persistence feature to Codex CLI that:
1. **Stores command history**: Commands are saved to
`~/.codex/history.json` and persist between CLI sessions.
2. **Navigates history**: Users can use the up/down arrow keys to
navigate through command history, similar to a traditional shell.
3. **Filters sensitive data**: Built-in regex patterns prevent commands
containing API keys, passwords, or tokens from being saved.
4. **Configurable**: Added configuration options for history size,
enabling/disabling history, and custom regex patterns for sensitive
content.
5. **New command**: Added `/clearhistory` command to clear command
history.
## Code Changes
- Added `src/utils/storage/command-history.ts` with functions for
history management
- Extended config system to support history settings
- Updated terminal input components to use persistent history
- Added help text for the new `/clearhistory` command
- Added CLAUDE.md file for guidance when working with the codebase
## Testing
- All tests are passing
- Core functionality works with both input components (standard and
multiline)
- History navigation behaves correctly at line boundaries with the
multiline editor
I had Codex read #182 and draft a PR to fix it. This is its suggested
approach. I've tested it and it works. It removes the purple `thinking
for 386s` type lines entirely, and replaces them with a single yellow
`thinking for #s` line:
```
thinking for 31s
╭────────────────────────────────────────╮
│( ● ) Thinking..
╰────────────────────────────────────────╯
```
prompt. I've been using it that way via `npm run dev`, and prefer it.
## What
Empty "reasoning" updates were showing up as blank lines in the terminal
chat history. We now short-circuit and return `null` whenever
`message.summary` is empty, so those no-ops are suppressed.
## How
- In `TerminalChatResponseReasoning`, return early if `message.summary`
is falsy or empty.
- In `TerminalMessageHistory`, drop any reasoning items whose
`summary.length === 0`.
- Swapped out the loose `any` cast for a safer `unknown`-based cast.
- Rolled back the temporary Vitest script hacks that were causing stack
overflows.
## Why
Cluttering the chat with empty lines was confusing; this change ensures
only real reasoning text is rendered.
Reference: openai/codex#182
---------
Co-authored-by: Thibault Sottiaux <tibo@openai.com>
