GitHub Repository

Your agent writes down why, in the repo, so everyone else's agents can find it without asking you.

0 starsTypeScript

Grepathy – Claude made a decision nobody approved

Name: Grepathy – Claude made a decision nobody approved
Availability: InStock
Author: evansjp

by evansjp·Jul 15, 2026·13 points·18 comments

Visit Project View on HN

AI Analysis

●●●BangerBig BrainSolve My ProblemZero to One

Grep for unapproved agent decisions before they break production, solving the vanished transcript problem.

Strengths

•Post-hoc transcript parsing avoids relying on agents to voluntarily log their own reasoning.
•PreToolUse hooks inject context automatically so future agents see past decisions without asking.
•Includes a blind eval report proving agents answer 'why' correctly only with this context.

Weaknesses

•Tightly coupled to Claude Code transcripts; unclear support for Cursor, Copilot, or other agents.
•Relies on team discipline to commit the generated markdown files before pushing code.

Post Description

hey HN - Claude pre-created users in Clerk with null emails/names as "guest users" on a contract job. Wasn't in any plan. The CTO asked why, and I didn't know! I didn't make that decision!

The reasoning was in a transcript on my laptop. Claude Code deletes those after 30 days by default. Two of my projects lost their whole history that way.

Grepathy distills transcripts locally into markdown committed with the code. Decisions only, never your messages, no server.

List every decision nobody approved:

grep -rn "agent-initiated" .ai/why/

Ran a blind eval before shipping, published it including the misses (REPORT.md). Agents with Grepathy answered the "why" questions right. Baseline agents made up confident wrong answers.