Gemini Models having problems with executing commands

[SuzayKid]

First off, sorry to open an issue when a similar one is available.. but i feel i can give more information on the issue and hence i opened a new one...

I am working with linux mint... even simple commands doesnt seem to work
Operating System: Linux Mint 22.1
Kernel: Linux 6.8.0-58-generic
Architecture: x86-64
Hardware Vendor: Apple Inc.
Hardware Model: MacBookAir7,2

open-codex -m gemini-2.0-flash-thinking-exp -a full-auto
╭──────────────────────────────────────────────────────────────╮
│ ● OpenAI Codex (research preview) v0.1.30 │
╰──────────────────────────────────────────────────────────────╯
╭──────────────────────────────────────────────────────────────╮
│ localhost session: ******************************* │
│ ↳ workdir: ~/Projects/******** │
│ ↳ model: gemini-2.0-flash-thinking-exp │
│ ↳ approval: full-auto │
╰──────────────────────────────────────────────────────────────╯
user
explain this codebase to me

command

$ 'ls -R'

command.stdout (code: 1, duration: 0s)

Error: spawn ls -R ENOENT

user
rewrite the test.txt file to 'i am from linux mint'

command

$ apply_patch '*** Begin Patch
*** Update File: test.txt
--- /dev/null
+++ b/test.txt
@@ -0,0 +1 @@
+i am from linux mint
*** End Patch'

command.stdout (code: 1, duration: 0s)

Invalid Context 0:
-- /dev/null

command

$ 'echo "i am from linux mint" > test.txt'

command.stdout (code: 1, duration: 0s)

Error: spawn echo "i am from linux mint" > test.txt ENOENT

codex
I attempted to rewrite the test.txt file for you using both the apply_patch
command and a standard shell echo command, but unfortunately, neither was
successful due to errors in the execution environment.

The apply_patch command failed with an "Invalid Context 0" error when trying
 to create or update the file.

The shell command echo "i am from linux mint" > test.txt failed because the
echo command was not found (ENOENT error).

It appears I am unable to modify or create files in this environment with
the current tools.

[Vlad1mir-D]

Try this fix: openai#695