Proof Portal

Project overview

Omarchy

ProbeLabs65 findings · 95 requirements

A 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.

Inbox
ReportR-1163GitHub issuePublicDecided· Out of scope

agents: Claude limits cache is shared across CLAUDE_CONFIG_DIR accounts

Reported by mikebronner · filed Aug 31, 2026 · open at source

Decided: Out of scope — decided by agents · verified 4 days ago.

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?
AreaOutside the audited area· Decided by agents · verified
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?
Decided by agents and verified by a second, independent agent — see the decision.

Original report

agents: Claude limits cache is shared across CLAUDE_CONFIG_DIR accountsrevision 1 · captured 2026-09-28 21:24 UTC via poll

System details

Omarchy 4.0.2-1 (stable), kernel 7.1.9-arch1-2, Hyprland 0.56.2, Wayland, AMD Ryzen AI 9 HX 470 / Radeon 890M, Mesa 1:26.2.1-1, 33GB RAM, Python 3.14.7, jq 1.8.2, Claude Code 2.1.251

What's wrong?

The claude collector reads the account from CLAUDE_CONFIG_DIR (omarchy-agent-usage-claude:39), and the plugin README documents that as the supported way to point at a non-default Claude directory. But the limits probe is cached at a fixed path that has no account in it:

probe_cache = cache_root() / "claude-limits.json"   # :800

So on a machine with two Claude accounts, whichever collector ran last owns the cache and the other one reads its numbers. Two ways it shows up:
1. Within PROBE_MIN_INTERVAL_SECONDS (15s) the second account reuses the first's limits outright.
2. Worse, the fallback is read with float("inf") max age, so if the second account's token is missing or expired the other account's percentages are shown as "the answer of record" — indefinitely, labelled as the wrong account.

Repro: sign into two accounts in separate config dirs, then

omarchy-agent-usage-claude --limits-only                       # account A
CLAUDE_CONFIG_DIR=~/.claude-other omarchy-agent-usage-claude --limits-only

Both print A's limits. I hit this with a 20x account reading 86% weekly and a 5x account reading 4%; the panel showed 86% for both.

Suggested fix: key the cache filename by the resolved config dir, the way the transcript scan cache already does with scan_cache_paths(). Workaround is a separate XDG_CACHE_HOME per account.

View raw text
### System details

Omarchy 4.0.2-1 (stable), kernel 7.1.9-arch1-2, Hyprland 0.56.2, Wayland, AMD Ryzen AI 9 HX 470 / Radeon 890M, Mesa 1:26.2.1-1, 33GB RAM, Python 3.14.7, jq 1.8.2, Claude Code 2.1.251

### What's wrong?

The claude collector reads the account from CLAUDE_CONFIG_DIR (omarchy-agent-usage-claude:39), and the plugin README documents that as the supported way to point at a non-default Claude directory. But the limits probe is cached at a fixed path that has no account in it:

```
probe_cache = cache_root() / "claude-limits.json"   # :800
```

So on a machine with two Claude accounts, whichever collector ran last owns the cache and the other one reads its numbers. Two ways it shows up:
1. Within PROBE_MIN_INTERVAL_SECONDS (15s) the second account reuses the first's limits outright.
2. Worse, the fallback is read with float("inf") max age, so if the second account's token is missing or expired the other account's percentages are shown as "the answer of record" — indefinitely, labelled as the wrong account.

Repro: sign into two accounts in separate config dirs, then

```
omarchy-agent-usage-claude --limits-only                       # account A
CLAUDE_CONFIG_DIR=~/.claude-other omarchy-agent-usage-claude --limits-only
```

Both print A's limits. I hit this with a 20x account reading 86% weekly and a 5x account reading 4%; the panel showed 86% for both.

Suggested fix: key the cache filename by the resolved config dir, the way the transcript scan cache already does with scan_cache_paths(). Workaround is a separate XDG_CACHE_HOME per account.

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.

  • Out of scopeDecided by agents · verified · 2026-10-09 10:18 UTC