Give coding agents compiler-accurate call graphs for refactors
Lantern Lattice gives coding agents call-graph tracing, impact analysis and symbol-based edits over MCP, such as renaming a function across 4 files in two calls; built largely with Claude Code plugins. ·
Done withClaude Code
- What they did
- Lantern Lattice runs as an MCP server that builds call graphs from each language's own tools, such as rust-analyzer, gopls and clangd. The agent uses trace to list a function's callers, then one replace_text call to rename it across files. Its tools replace Grep, Read, Edit and Write, which keeps the graph current. Each non-Rust language was built mostly through its own Claude Code plugin.
- How it went
- Renaming a function took two calls: 550 bytes for trace, 375 for the edit across 4 files. Grep plus reading the files was 26 lines in 23 files and 51 KB. Uncertain calls are left out.
- Worth knowing
- Only tested with Claude Code so far; configs for other agents are untested. Rust queries on 20k symbols take 0.01 to 0.02 s, and the first answer after an edit about 1 s.
Try it yourself with Claude Code
In my repository at [path or repo name], I want to rename the function [function name] to [new name] safely. First trace everything that calls it and list the files and call sites affected, then make the rename across all of them, including tests and docs. Run [test command] afterward and stop once the tests pass, giving me a summary of every file changed.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
Source: HN · Nov 5, 2026
Related use cases
Turn tickets into merged pull requests
Pulls tasks from GitHub Issues or Linear, then runs Claude Code or Codex in isolated Kubernetes pods to carry each one through to a PR.
Catch up on two years of overdue Dutch taxes
Claude Code helped a person almost two years behind on taxes untangle expenses, look up receipts and invoices, and navigate the Dutch tax system.
Run Claude Code on local models via Ollama
A guide to setting up Claude Code in IntelliJ and connecting it to local models through Ollama for coding work.
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Resolve a Slack-reported issue with a coding agent
A Codex agent resolved an issue passed to it in about 40 minutes, and an ops agent verified the fix back in Slack.
Turn a website into a mobile app
The author used Replit Agent 4 over about two days and roughly $40 to turn a website into a mobile app with design, preview, and App Store prep, then submitted it for review.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.