James Washington
writing about archive utils free game ideas replays
Document

jev client and whitepaper

A small shared helper for a decision-making AI model, plus a report evaluating how well it works.

built Sep 2026 Letting AI make calls, safely

how it works

Stdlib-only Python client with canary monitor; whitepaper evaluates Jev against a human review ledger.

Document: A reference or runbook.

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

evalstypesafe

related

  • agent-merge: Jev-gated self-merge for agent branches Automation The one sanctioned way an AI may merge its own work: a script that reviews it first and refuses if it didn't show its work.
  • Code-owns-workflow, models-supply-judgment outreach pipeline Automation A program that researches leads and writes draft sales emails for a person to approve, never sending anything itself.
  • Company policies for an agent-run business (autonomy, delegation, spending) Document A written rulebook that lets AI agents run a small business's day-to-day without asking the owner, covering honesty, what they may decide alone, and how much they may spend.
  • Wiki SCHEMA.md: a contract any AI can follow Document A one-page rulebook that tells any AI how to add to and maintain the personal wiki.
  • Jev shadow scorer for recruiter mail Automation Tests an AI scoring of recruiter emails quietly, logging results without letting it decide anything yet.
← everything I've built
© 2026 James Washington
github twitter rss