Extracts learning signals from chat sessions and persists them to configuration files.
npx skills add https://github.com/cuipengfei/prompts --skill auto-extract
Please help me install this Agent Skill. Skill: auto-extract Source: https://github.com/cuipengfei/prompts/tree/main/plugins/session-learn/skills/auto-extract Command: npx skills add https://github.com/cuipengfei/prompts --skill auto-extract
This Skill reduces repeated trial-and-error by automatically analyzing your current or historical AI coding sessions and extracting high-signal learning to persist as configuration knowledge.
Ask your AI to run auto-extract for the current conversation so it can extract learning signals and propose where to persist them, then confirm the suggested writes.
High-intent search queries and answers about installing and using this skill.
To save AI chat learnings into durable configuration rules, you need a tool that extracts learning signals from conversations and persists them into appropriate configuration layers. This process analyzes session data for progress wins, failures, and user corrections, then writes them to configuration files like CLAUDE.md or AGENTS.md after probing available backends and confirming with you.
The best way to extract configuration knowledge from historical agent logs is to use batch analysis triggered by an explicit prefix. This approach identifies high-signal learnings across multiple sessions, including failure points and inferred preferences, and routes them into the correct storage layer with de-duplication before safe persistence.
Agent memory routing works by directing each extracted learning signal into one of four layers: L1 user configuration, L2 project root, L3 submodule, or L4 situation memory. The system performs cross-layer de-duplication and requires user confirmation before writing to ensure the most semantically appropriate target is updated.
Yes, you can use it to update CLAUDE.md and AGENTS.md files automatically. The system probes your environment to detect available agent configuration files and memory backends without hardcoding paths, then creates or updates the most semantically appropriate targets after you confirm the suggested writes.
You can reduce repeated trial-and-error in AI coding sessions by automatically analyzing your current or historical sessions and extracting high-signal learning to persist as configuration knowledge. This ensures that progress wins, failure points, and explicit corrections are captured and applied to future interactions.
A limitation of using session learning for prompt engineering configuration management is that safe persistence requires user-gated updates, meaning you must manually confirm suggested writes. Additionally, it relies on robust environment probing to detect available files and memory backends, which may limit automation if your setup is non-standard.
strawgateAnalyze agent-user conversation transcripts to identify skill gaps and failure patterns.
oimiragieoSynchronize memory patterns across CLAUDE.md and the Memory Tool.
RobbyMo81Coordinate multi-agent workflows in OpenProse with sessions, imports, and state backends for parallel execution and persistence.
overcut-aiDocument Overcloud agent tool categories, descriptions, and built-in type matrices.
