Proof Portal
Omarchy
ProbeLabs65 findings · 95 requirementsA proof layer — requirements, tests and verified fixes — for two of Omarchy's subsystems: the application menu (launcher scripts, QML model, JSONC config, search and selection) and the lock screen (lock scripts, QML, PAM authentication). Scope is deliberately limited to those components of omacom/omarchy; the rest of the distribution is not covered.
Agents widget: Codex limits intermittently fail with "account/read" (select/readline race)
Reported by JoacoRivera · filed Sep 24, 2026Closed upstream on 2026-10-04 by the bot omarchybot, as a duplicate of #13080 (R-150). GitHub close reason: completed.
At a glance
- 1 · What is claimed?
- Only the report text below — no validator has restated it as checkable claims yet.
- 2 · What does it affect?
- Not yet mapped to any requirement or component.
- 3 · Is it reproducible?
- No validation has been run yet.
- 4 · What remains unclear?
- Not validated yet: no claims, no reproducer.
- 5 · What happens next?
- Our staff decide whether the report is valid. Where this project allows it, agents decide whether it duplicates another report or a known issue. No decision has been recorded yet.
Original report
System details
- Omarchy 4.0.4-1
- CPU: AMD Ryzen 5 5500U
- GPU: AMD Radeon (Lucienne, integrated)
- codex-cli 0.156.1
What happens
The Codex tab of the agents widget intermittently shows "Codex limits unavailable" with account/read as the help text, even though Codex is logged in and the app-server answers in under a second. It tends to stick after a reboot (nothing refreshes the record again for refreshIntervalSec), and opening a Codex session "fixes" it only because it leads to another refresh that happens to win.
Steps to reproduce
for i in $(seq 8); do omarchy-agent-usage-codex --limits-only | jq -r '.usageStatusText + " | " + .authHelpText'; doneOn my machine about 1 in 4 runs prints Codex limits unavailable | account/read; the rest succeed. There is no network or auth problem: a blocking-read client against the same codex app-server succeeds every time (account/read ≈ 0.4 s, account/rateLimits/read ≈ 0.7 s).
Cause
rpc_request() in bin/omarchy-agent-usage-codex waits with select.select([proc.stdout], ...) on the fd, then reads with proc.stdout.readline() on a buffered text stream:
ready, _, _ = select.select([proc.stdout], [], [], 0.25)
if not ready:
continue
line = proc.stdout.readline()Right after initialize/initialized, the app-server emits notifications (remoteControl/status/changed, account/updated). When one of them arrives in the same chunk as the account/read reply, readline() pulls the whole chunk into Python's buffer and returns only the first line. The reply is already buffered, so the fd never becomes readable again, select() keeps timing out, and TimeoutError("account/read") is raised. Whether it fails depends only on how the output happens to be chunked, which is why it is intermittent.
Suggested fix
Either check the buffer before selecting, or read the fd without buffering and split lines yourself, for example:
buf = b""
while True:
while b"\n" in buf:
line, buf = buf.split(b"\n", 1)
... # parse, return if id matches
if not select.select([proc.stdout], [], [], remaining)[0]:
raise TimeoutError(method)
chunk = os.read(proc.stdout.fileno(), 65536)
if not chunk:
break
buf += chunk(The buffer must persist across rpc_request() calls, since a reply can arrive together with the previous one.) With that change I get 10 out of 10 successful probes locally.
View raw text
### System details
- Omarchy 4.0.4-1
- CPU: AMD Ryzen 5 5500U
- GPU: AMD Radeon (Lucienne, integrated)
- codex-cli 0.156.1
### What happens
The Codex tab of the agents widget intermittently shows **"Codex limits unavailable"** with `account/read` as the help text, even though Codex is logged in and the app-server answers in under a second. It tends to stick after a reboot (nothing refreshes the record again for `refreshIntervalSec`), and opening a Codex session "fixes" it only because it leads to another refresh that happens to win.
### Steps to reproduce
```bash
for i in $(seq 8); do omarchy-agent-usage-codex --limits-only | jq -r '.usageStatusText + " | " + .authHelpText'; done
```
On my machine about 1 in 4 runs prints `Codex limits unavailable | account/read`; the rest succeed. There is no network or auth problem: a blocking-read client against the same `codex app-server` succeeds every time (`account/read` ≈ 0.4 s, `account/rateLimits/read` ≈ 0.7 s).
### Cause
`rpc_request()` in `bin/omarchy-agent-usage-codex` waits with `select.select([proc.stdout], ...)` on the fd, then reads with `proc.stdout.readline()` on a buffered text stream:
```python
ready, _, _ = select.select([proc.stdout], [], [], 0.25)
if not ready:
continue
line = proc.stdout.readline()
```
Right after `initialize`/`initialized`, the app-server emits notifications (`remoteControl/status/changed`, `account/updated`). When one of them arrives in the same chunk as the `account/read` reply, `readline()` pulls the whole chunk into Python's buffer and returns only the first line. The reply is already buffered, so the fd never becomes readable again, `select()` keeps timing out, and `TimeoutError("account/read")` is raised. Whether it fails depends only on how the output happens to be chunked, which is why it is intermittent.
### Suggested fix
Either check the buffer before selecting, or read the fd without buffering and split lines yourself, for example:
```python
buf = b""
while True:
while b"\n" in buf:
line, buf = buf.split(b"\n", 1)
... # parse, return if id matches
if not select.select([proc.stdout], [], [], remaining)[0]:
raise TimeoutError(method)
chunk = os.read(proc.stdout.fileno(), 65536)
if not chunk:
break
buf += chunk
```
(The buffer must persist across `rpc_request()` calls, since a reply can arrive together with the previous one.) With that change I get 10 out of 10 successful probes locally.
Validation
Not checked yet. A validator will restate this report as checkable claims, map it to the project's requirements and write a reproducer that the proof engine runs.
Linked
Not linked to any known issue, other report or change yet.
Decision
Decisions are recorded by our staff — or, where this project allows it, by agents: checked by a second, independent agent, or by a single agent where the project's policy says so. Any decision reached so far is shown below.
No decision has been recorded yet.