Agent Memory Is Not a Database: Start with an Event Log
I break agent memory into writing, storage, retrieval, assembly, and forgetting, then outline a minimal implementation path that starts with a durable event log.
Category
Learning
The Agent Extension Stack: Skills, Plugins, MCP, and Tools
A Skill teaches an agent how to work, a Tool lets it act, MCP connects it to external capabilities, and a Plugin packages the pieces for installation.
Reading World Models Through 'LLMs Can't Jump'
A reading note on abductive reasoning, interactive simulation, and what current world models can and cannot do.
What Pi Coding Agent Keeps in the Core
A source-based comparison of Pi, Codex, and OpenCode shows how a minimal coding harness leaves workflow policy and isolation to the user.
Inside Matt Pocock's Skills: Where the Control Lives
A commit-pinned reading of the Matt Pocock Skills repository shows how its invocation rules keep workflow direction with the human while the host retains tools and permissions.
From Superpowers to Skills: What Changed with GPT-5.6
After moving to GPT-5.6, I began choosing smaller agent workflows and paying more attention to the loop that decides what happens next.
Automating a Mobile App with an Android Emulator
I used Python, an Android emulator, and OCR to automate a repetitive in-app card-spinning workflow.
How to Write Better Prompts for LLMs
Four practical ways to give language models clearer instructions, better context, and smaller tasks.
怎樣才能寫好提示詞？
四個實用方法：講清楚任務、補充背景、要求檢查，再把大任務拆小。