You've hit your session limit · resets 1:30pm (Europe/Prague)

**Bug Description**
Additional context for my previous usage-limit report:

I want to clarify that today's workload was not unusually heavy compared with my normal Claude usage.

My computer was asleep overnight. After waking it this morning, I generated roughly 14 reports on claude.ai and then used Claude Code with Claude-in-Chrome/Edge to transfer those reports into a web portal. This is a fairly routine workload for me and substantially lighter than some previous days where I did not exhaust the 5-hour limit nearly this quickly.

I did have several Claude Code PowerShell sessions left open from yesterday, but they were idle and I was not interacting with them. The machine was also asleep overnight.

What concerns me is the discrepancy between the amount of work performed and the quota consumption. One local Claude Code session showed approximately:

55m49s wall time
72 requests
73.4k output tokens
9.1M cache-read tokens
228.3k cache-write tokens
97% of input tokens served from cache
~$10.54 API-equivalent cost

Yet my 5-hour allowance reached 100%.

The local usage diagnostic also reports that 88% of the previous 24h usage involved subagent-heavy sessions and 82% involved >150k context, but much of that activity was from yesterday, not from the current 5-hour window.

The important point is that this appears to be a significant regression relative to my established workload. I routinely perform comparable or heavier workflows without exhausting the limit this rapidly.

Could you please check the account-side usage/metering logs and confirm whether the consumption is expected, particularly for Fable 5.1 and Claude-in-Chrome sessions?

I can provide additional screenshots/session diagnostics if useful.

**Environment Info**
- Platform: win32
- Terminal: windows-terminal
- Version: 2.1.268
- Feedback ID: 4305b13c-3844-450e-9f3f-1fd68d323c3b

**Errors**
```json
[]
```

No repository files, tests, or code entry points are mentioned. Start by reviewing the reported 5-hour quota, session diagnostics, reset timing, and account-side usage or metering logs; compare the 72 requests and token counts with the current window and Claude-in-Chrome activity. Done means confirming whether consumption is expected or identifying a metering discrepancy with supporting diagnostics.

索引モデルが issue の本文から書いたものです。