I Pointed REA at a 100 MB Electron App. Here’s What It Found and the Two Preload Files It Missed.

Description and use

REA’s static analysis is fast and largely accurate on a 100 MB Electron bundle, but it fails to resolve dynamic preload paths and exposed APIs, and drops many graph connections due to budget limits. On a toy app it matches ground truth and can trace dynamic IPC. Agent cost is significant (~1.95 MB JSON).

Problem

Assessing the effectiveness of REA’s static JavaScript analysis on a large Electron application and a controlled toy app, and evaluating agent resource costs.

Outcome

REA’s static analysis is fast and largely accurate on a 100 MB Electron bundle, but it fails to resolve dynamic preload paths and exposed APIs, and drops many graph connections due to budget limits. On a toy app it matches ground truth and can trace dynamic IPC. Agent cost is significant (~1.95 MB JSON).

Workflow

• Install Node 22.20 and REA via npm
• Run rea doctor to check environment
• Set up REA with npx rea setup --client claude_code --dry-run and --yes
• Analyze the 100 MB bundle with npx rea analyze-javascript-application app.asar --format json
• Compare REA’s summary to grep counts
• Analyze the toy app before and after minification
• Use trace-application-feature to trace IPC channels
• Measure agent tool list size and latency

Evidence

REA read 459 files (24.2 MB of text), parsed 421 JavaScript files, walked 3,114,427 AST nodes and had zero parse failures.

Source

REA said 0 preload entrypoints. The app has two preload files, preload/index.cjs and preload/plugin-panel.js.

Source

The tracer is the highlight. I fed the evidence back in with trace-application-feature and a seed of channel: "notes:read". It returned the chain I wanted, on the minified bundle: preload.js invokes the channel notes:read, and the channel is handled by the notes:read handler.

Source

REA exposes 170 MCP tools. I opened a stdio session