Loading page...Загружаем страницу...正在加载页面...
AtlasRepo

Six agents were running and I could not tell you what any of them did

Description and use

runlanes automatically aggregates logs from six AI coders, showing token spend, concurrency, and simultaneous file edits, simplifying post‑run audit.

Problem

When multiple coding agents run concurrently, it is hard to determine which agent performed which actions, how much token cost each incurred, and whether they edited the same files.

Outcome

runlanes provides a transparent, post‑run audit of multiple AI coding agents, revealing token usage, concurrency, and file edit overlaps without requiring instrumentation or prior setup.

Workflow

  • 1. Install runlanes via npm or npx
  • 2. Run npx runlanes in the project root
  • 3. The tool scans the home directory for agent transcripts in six formats
  • 4. It aggregates data into tabs: Now (live sessions), Graph (shared files), History (past runs), Docs (project instructions)
  • 5. Review token spend, concurrency, and file edit overlaps
  • 6. Export the console as a self‑contained HTML with runlanes --export for sharing or CI validation

Evidence

runlanes reads local files, makes no outbound network calls, and binds to 127.0.0.1.

Source

The console shows a dash. It would be trivially easy to show 0 there.

Source

  • Record date: 2026-09-09T00:31:15.101Z