Pinned Posts

Deep Dive into Claude Code Harness - Chapter 1: Memory Management

What makes the Claude Code harness best-in-class? Let's dive into its leaked source code to learn about the rigorous engineering within. Turns out, its memory management is remarkably similar to the wiki system described in my previous post.

Introducting Wiki-Skills: Agent Skill for Managing Markdown Vaults

Kaparthy shared a viral note about a building an LLM-managed 'personal wiki' - which I interpret as a specific instance of a locally stored markdown vault. Here, I present Wiki Skills - a set of agentic Skills that captures the process of creating, querying, and updating this knowledge vault.

Autoresearch: Autonomous Hill-Climbing for Any Optimizable System

Run optimization experiments while you sleep

Coffee Water Composition

TL;DR: Boston’s tap water is terrible for making high quality coffee (wrt aroma and taste). No need to start at zero with deionized water – spike your water to get better results. In these notes, I aim to derive a recipe for optimizing my flat whites using Boston’s (filtered) tap water.

A Review on Agentic Memory

Managing context for long-running agents, across sessions, or swarms of agents, remains an active area of research and development. This is a survey of the current landscaspe in the world of Agentic memory.

Python Tooling

An index of relevant Python tooling references in this blog

Tower Defense

A Javascript version of the original Flash-based Tower Defense, circa 2008

Coffee Water Composition

TL;DR: Boston’s tap water is terrible for making high quality coffee (wrt aroma and taste). No need to start at zero with deionized water – spike your water to get better results. In these notes, I aim to derive a recipe for optimizing my flat whites using Boston’s (filtered) tap water.

Read More

A Review on Agentic Memory

Managing context for long-running agents, across sessions, or swarms of agents, remains an active area of research and development. This is a survey of the current landscaspe in the world of Agentic memory.
Read More

Save ADRs, Discard Design and Implementation Docs

After adoping the obra/superpowers workflow, design docs became an important part of my coding process. They’re detailed and verbose - containing code snippets, specific tasks and implementation procedures, and tests - perfect for handing off to subagents for implementation. But after the feature is built, I couldn’t decide whether to include them as part of my git history or to discard them. 95% of their content will probably never be consumed by either me or LLMs ever again, but they do contain little snippets of historical context that may be important for future architectural decisions.

Read More