analyze_dead_code
Analyze a completed scan to identify dead JavaScript functions.
Reads V8 coverage data collected during the scan, merges execution counts across passes, and reports every function with zero executions.
LABSCO SUMMARY
Deletion backed by coverage, with a way back.
The evidence for removing a function is that it never ran while a browser agent used the site, and risk scores plus pre-patch snapshots mean a wrong call is reversible. Coverage only proves what the crawl exercised, which is what the verification replay is for.
What it is
An MCP server for Cursor that finds dead JavaScript by driving a real browser over your site, collecting V8 precise coverage, and turning zero-execution functions into removal patches.
What you get
- One call that runs the whole pipeline and returns file paths, line numbers, risk scores and removal recommendations (find_dead_code)
- Scan and analysis as separate steps when one crawl should feed several passes (scan_website, analyze_dead_code)
- Unified-diff patches generated at a conservative, balanced or aggressive setting, and readable before they are applied (generate_patches, get_patch_diff)
- Patches applied to disk with pre-patch snapshots kept, and a rollback that restores the originals (apply_patch, rollback_patch)
- A verification replay that re-runs the browser agent over the same interactions after patching (verify_patch)
- Recent scans listed with their status, and one scan's detail (list_scans, get_scan_status)
Requirements
Python 3.11+, Playwright Chromium installed with playwright install chromium, and a Gemini API key, which the first run writes into ~/.cursor/mcp.json.
Setup effort
One command plus a key — pipx run codereaper, then supply credentials
USE CASES
01
Finding out how much of a bundle never runs
Point it at the site and read back the dead functions with file paths and line numbers.
02
Cutting dead code out of a legacy front end
Generate the diff, read it, then apply it to the source files.
03
Checking a removal did not break the UI
Replay the original interactions after the patch, and restore the files if something regressed.
04
Re-analysing an old crawl at a bolder setting
Reuse the scan instead of crawling again, and regenerate patches from it.
05
Watching a long scan
Start the crawl on its own and check its status while it runs.
10 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
apply_patch
Apply a generated patch to the source files on disk.
Stores pre-patch snapshots so the change can be rolled back later. Review the diff (get_patch_diff) before applying.
find_dead_code
Scan a website and find dead JavaScript code using an AI browser agent.
This is the primary tool. It launches a Chromium browser with an AI agent that autonomously explores the site (clicks buttons, navigates routes, opens modals, fills forms, triggers keyboard shortcuts) while V8 precise coverage tracks which JavaScript functions actually execute.
Functions that receive zero executions across all exploration passes are reported as dead-code candidates with risk scores and removal recommendations.
generate_patches
Generate unified-diff patches that remove dead code.
Creates a patch set based on the analysis results for the given scan. Each affected file gets a unified diff that replaces dead functions with a removal comment.
get_patch_diff
Retrieve the combined unified diff for a generated patch.
get_scan_status
Get detailed status and metadata for a specific scan.
list_scans
List recent scans with their status.
Returns: JSON array of scan records (most recent first, up to 20).
rollback_patch
Rollback a previously applied patch, restoring the original files.
scan_website
Launch the AI browser agent to explore a website and collect V8 coverage.
Use this when you want to run the scan separately from the analysis (e.g. to inspect scan status first). After it completes, pass the returned scan_id to analyze_dead_code().
verify_patch
Verify an applied patch by replaying the original interactions.
Re-runs the browser agent with the same interaction plan used during the scan and checks for regressions: new console errors, missing elements, and coverage drops.
BEFORE YOU INSTALL
What to know first
Tools you get
10 tools when we connected.
Requires
- Python 3.11+
- Playwright Chromium (playwright install chromium)
- A Gemini API key, which codereaper asks for on first run (or an OpenAI or Anthropic key if you change providers)
CONTEXT COST
Light Standing cost
Rated by the size of the 10 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
Setup
TERMINAL
pip install codereaper
playwright install chromiumRun codereaper once: it asks for your Gemini API key and adds itself, with that key, to your global ~/.cursor/mcp.json
Restart Cursor
You do not set GEMINI_API_KEY by hand: the first time you run codereaper it asks for your Gemini key and saves it in your global ~/.cursor/mcp.json, which is where Cursor passes it to the server from.
pipx
TERMINAL
pipx run codereaperYou do not set GEMINI_API_KEY by hand: the first time you run codereaper it asks for your Gemini key and saves it in your global ~/.cursor/mcp.json, which is where Cursor passes it to the server from.
