[BUG] /context shows identical token count for all MCP tools

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

All MCP tools in the /context output show exactly the same token count (290 tokens each), regardless of their actual size. Tools with
extensive parameter schemas and descriptions (e.g. find_symbol) show the same value as minimal tools (e.g. list_memories).

Observed with Serena (23 tools) and Context7 (2 tools). Not sure if this affects all MCP servers or is specific to certain ones.

MCP tools · /mcp
└ mcp__context7__resolve-library-id: 290 tokens
└ mcp__context7__query-docs: 290 tokens
└ mcp__serena__list_dir: 290 tokens
└ mcp__serena__find_symbol: 290 tokens
└ mcp__serena__write_memory: 290 tokens
...
(all 25 tools show exactly 290 tokens)

Total: 25 × 290 = 7,250 — close to the summary "MCP tools: 6.7k tokens", suggesting even distribution rather than actual per-tool
measurement.

What Should Happen?

Each tool should show its actual token cost based on its description and parameter schema.

Error Messages/Logs

Steps to Reproduce

type /context when you have multiple MCP methods.

Claude Model

None

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.50 (Claude Code)

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_