James Washington
writing about archive utils free game ideas replays
Agent hook

PostToolUse file soft-claim hook for agent collision detection

Every time an AI session edits a file, it quietly tells the dashboard, so two sessions editing the same file get flagged.

built Jun 2026 updated Oct 2026 Upgrading the AI's workbench

how it works

Claude Code PostToolUse(Edit|Write|...) hook reads the tool payload from stdin, posts a short-lived soft claim; fail-open with a short timeout so the hub being down never blocks work.

Agent hook: A script that runs automatically at a moment in an AI session (before a command, when it stops) to steer or guard it.

The code for this one lives in a private repo. This page explains the idea so you can build your own.

claude-codecoordinationfail-openhooksmulti-agent

related

  • Mission Control presence/telemetry hooks Agent hook Every AI session automatically shows up live on the operations dashboard: when it starts, works, edits files and stops.
  • Mini build guard Agent hook Prevents AI assistants from running heavy builds on the home server and points them at the laptop instead.
  • One-shot scheduled agent follow-up Automation A one-time scheduled AI run that checked on phone calls it had placed and reported the outcome, then deleted itself.
  • /claude-replay: session to interactive replay post Slash command Turns an AI coding session into an interactive replay published on the blog.
  • commit_gate: no wandering off with a dirty tree Agent hook Stops an AI from finishing a task while leaving its changes uncommitted.
← everything I've built
© 2026 James Washington
github twitter rss