Claude Code's built-in Edit tool uses string matching. To change five lines, the model echoes back those exact lines as `old_string`, then provides the replacement. That echoed text is pure overhead (it's already in the file) and it's spending output tokens, the most expensive token class, just to say "I mean this part."

For a typical 15-line edit, that's ~200 wasted output tokens. Do a few dozen edits in a session and you're burning real money on text the model already knows is there. Worse, if `old_string` appears more than once in the file, the edit fails and the model has to pad extra context lines until the match is unique.

I built an MCP plugin that replaces string matching with line-range references and hash verification. The model says which lines to replace, proves it read them correctly with a checksum, and provides only the new content. A 15-line edit goes from ~470 output tokens to ~263. That's a 44% reduction. If the file changed since the last read (you saved in your editor, another tool touched it), the hash check catches it instead of silently applying a stale edit.

Install is two commands:

    /plugin marketplace add rjkaes/trueline-mcp
    /plugin install trueline-mcp@trueline-mcp

Session hooks automatically redirect the agent to use trueline tools.

Inspired by Can Boluk's "The Harness Problem" and Seth Livingston's vscode-hashline-edit-tool for VS Code.

Blog post with more detail: https://www.wormbytes.ca/2026/03/04/trueline-mcp-announcemen...