Show HN: ProvLedger – stops coding agents repeating mistakes

Category: ai-ml

Tags: provenance, decision-tracking, coding-agents, data-lineage, knowledge-management

Score: 8.0/10 (Innovation: 8, Technical: 8, Documentation: 9, Utility: 7)

provLedger is a Claude Code plugin that captures and organizes the rationale behind data and code changes, linking decisions to specific project entities (functions, columns, tasks) rather than sessions. It provides a queryable ledger with cited answers to 'why' questions, preserving history even for deleted items, and surfaces past decisions to coding agents to prevent repeated mistakes.

Target audience: data engineers, backend developers, AI coding agent users, engineering teams

Repository: https://github.com/yizhao95/prov_ledger · Python · NOASSERTION · 28 stars

View on Hacker News