← All topics

AI agents and coding workflows

A reading path through Shivam Agarwal’s writing on agent memory, Claude Code, test-driven development, and retrieval-augmented generation.

This collection brings together my writing on AI agents: how they use context and memory, how coding agents fit into development workflows, and how retrieval connects AI with information. The linked writing pages contain short introductions; the complete articles are on Medium.

Where should I start?

Is there code to explore?

The Claude Code × TDD project links to an example Python authentication service built around test-driven pair programming. It includes email/password registration, bcrypt password hashing, and in-memory or PostgreSQL storage options.

For the retrieval side of AI systems, explore search and Elasticsearch. For my professional background, see technology strategy and engineering work.

Writing in this collection

tkr: A Git-Native Task Tracker Built for AI Coding Agents

A task tracker designed around Git and the workflows of AI coding agents.

June 28, 2026 · On Medium

Claude Code: Using Claude Code with GLM-5 and MiniMax M2.5

Using alternative models with Claude Code.

February 23, 2026 · On Medium

Claude Code: TDD Pair Programming Sub-agents with minimal supervision

Pair programming with sub-agents and a test-driven development workflow.

February 15, 2026 · On Medium

Agentic AI: Pi — Anatomy of a minimal coding agent powering OpenClaw

A look inside Pi, the minimal coding agent powering OpenClaw.

February 6, 2026 · On Medium

Agentic AI: OpenClaw/MoltBot/ClawdBot’s Memory Architecture Explained

How OpenClaw, MoltBot, and ClawdBot organize agent memory.

February 5, 2026 · On Medium

RAG: Giving Your AI a Library Card

An introduction to retrieval-augmented generation.

August 31, 2025 · On Medium