[Bug] Session token limit reached prematurely after upgrade to v2.0.64
Bug Description claude codeの2.0.64にあげてからcurrent sessionの上限に達するのが異様に早くなった。 同様の報告をしている人もいるので、調査をお願いしたい。 https://x.com/benvargas/status/1998908915229376901
Environment Info
- Platform: darwin
- Terminal: WarpTerminal
- Version: 2.0.64
- Feedback ID: f34cf6d2-d37c-46d0-8aa4-52cb48c51f6b
Errors
[{"error":"Error: unable to get local issuer certificate\n at gBA.from (file:///Users/csap8163/.volta/tools/image/node/22.19.0/lib/node_modules/@anthropic-ai/claude-code/cli.js:173:14882)\n at RC.<anonymous> (file:///Users/csap8163/.volta/tools/image/node/22.19.0/lib/node_modules/@anthropic-ai/claude-code/cli.js:187:10016)\n at RC.emit (node:events:531:35)\n at RC.emit (node:domain:489:12)\n at bJ1.<computed> (file:///Users/csap8163/.volta/tools/image/node/22.19.0/lib/node_modules/@anthropic-ai/claude-code/cli.js:186:3432)\n at ClientRequest.emit (node:events:531:35)\n at ClientRequest.emit (node:domain:489:12)\n at emitErrorEvent (node:_http_client:105:11)\n at TLSSocket.socketErrorListener (node:_http_client:518:5)\n at TLSSocket.emit (node:events:519:28)\n at UWA.request (file:///Users/csap8163/.volta/tools/image/node/22.19.0/lib/node_modules/@anthropic-ai/claude-code/cli.js:189:2130)\n at process.processTicksAndRejections (node:internal/process/task_queues:105:5)\n at async GG0 (file:///Users/csap8163/.volta/tools/image/node/22.19.0/lib/node_modules/@anthropic-ai/claude-code/cli.js:4548:3922)","timestamp":"2025-12-11T01:30:22.947Z"}]
Found 3 possible duplicate issues:
- https://github.com/anthropics/claude-code/issues/13552
- https://github.com/anthropics/claude-code/issues/13551
- https://github.com/anthropics/claude-code/issues/13536
This issue will be automatically closed as a duplicate in 3 days.
- If your issue is a duplicate, please close it and 👍 the existing issue instead
- To prevent auto-closure, add a comment or 👎 this comment
🤖 Generated with Claude Code
+1
+1
+1
+1
+1 same for me.
+1 on Max 10x plan
+1
⭐ TIPS:
I've been browsing similar area:cost issues re: Claude this morning and have gathered a few tips in one place. These don't fix the issue entirely but they've helped cut down my usage in the interim: hope these help someone!
- direct Claude to not use Task feature, and to use utility tools for common tasks. Add
–no-agent flagto your prompt as well (several users reporting that Claude is recruiting subagents without disclosing this to the user)
- ie “ --no-agent for the remainder of our work in this session. DO NOT use the Task tool. Use Grep to find files, Read to check them, and Edit to fix them directly."
- use
/configin the CLI to disable auto-compact, as this is another source of token overconsumption
If anyone else finds ways to reduce token consumption while we wait, feel free to share! 👍
⭐ TIPS: I've been browsing similar
area:costissues re: Claude this morning and have gathered a few tips in one place. These don't fix the issue entirely but they've helped cut down my usage in the interim: hope these help someone!1. direct Claude to _not_ use Task feature, and to use utility tools for common tasks. Add `–no-agent flag` to your prompt as well (several users reporting that Claude is recruiting subagents without disclosing this to the user) * ie “ --no-agent for the remainder of our work in this session. DO NOT use the Task tool. Use Grep to find files, Read to check them, and Edit to fix them directly." 2. use `/config` in the CLI to disable auto-compact, as this is another source of token overconsumptionIf anyone else finds ways to reduce token consumption while we wait, feel free to share! 👍
One more thing I noticed in practice: MCP servers can consume a noticeable amount of tokens simply by being connected, even if you’re not actively using them.
Because of that, it’s worth disabling MCPs that aren’t needed for the current task or project. In one of my projects, doing this alone reduced token usage by about 17%.
Also, it helps to periodically check /context to see where the token budget is actually going.
One more thing I noticed in practice: MCP servers can consume a noticeable amount of tokens simply by being connected, even if you’re not actively using them.
Because of that, it’s worth disabling MCPs that aren’t needed for the current task or project. In one of my projects, doing this alone reduced token usage by about 17%.
Also, it helps to periodically check
/contextto see where the token budget is actually going.
Thank you. That's helpful. But I'd still like to wait for an answer to the main question. Don't you agree?