Grepathy: Logging Agent Decisions
evansjp · hn · 2026-07-15
Grepathy: Decision Logging for Coding Agents
During a contract project, the author discovered that Claude would autonomously pre-create "guest users" with null emails/names—an unapproved decision. Worse, Claude Code deletes transcripts by default after 30 days, causing the complete loss of historical context across two projects.
To fix this, the author built Grepathy: it compresses transcripts into Markdown locally and commits them alongside the code, logging only "decisions" rather than full conversations.
Usage
- Save agent decision descriptions in the repo (e.g., path: .ai/why/)
- Use grep -rn "agent-initiated" .ai/why/ to find actions "initiated by the agent but unapproved by humans"
Results
- Conducted a blind eval before release and publicly shared the report and missed items
- Agents equipped with Grepathy can correctly answer follow-up questions like "why was this done"
- Baseline agents would confidently hallucinate incorrect answers
More from coding & agent
- A better path to agent autonomy is running waves, finding friction, and iterating — JnBrymn · 2026-07-22
- Coding agents are heading toward an AI-writes, AI-reviews, human-approves workflow — aftahi_ai · 2026-07-22
- oMLX 0.5.2 adds Mac menu-bar stats, low-bit decode kernels, and faster downloads — awnihannun · 2026-07-22
- GitHub review bot hits its PR limit and forces a 39-minute cooldown — DanielLockyer · 2026-07-22
- Max reasoning effort appears to be mobile-only in Codex Remote, not desktop — GabGarrett · 2026-07-22
- A Reddit demo argues online stores should expose carts and pricing through MCP — gelembjuk · 2026-07-22