Show HN: Standing Questions – agent memory that stores questions, not answers

Category: library

Tags: ai-agent, memory-pattern, developer-tools

Score: 6.8/10 (Innovation: 7, Technical: 6, Documentation: 8, Utility: 6)

Standing Questions is a memory pattern for AI coding agents that stores persistent questions instead of mutable answers, forcing agents to re-derive facts from the current codebase each session. It ships with JSON schemas, a seed question set, re-derivation logging, and Python metrics to detect when agents are going through the motions. The project is interesting for its inversion of typical memory design and its pragmatic focus on failure-mode detection through simple, verifiable artifacts.

Target audience: backend devs, AI/ML engineers, devops

Repository: https://github.com/Rocco-alt/standing-questions · Python · MIT

View on Hacker News