Seevali Rathnayake

Numbers

performance bottlenecks catalogued and targeted

.NET projects unified into a single monorepo

BUILD May 25, 2026 2 of 2

Watching the Loop Forge Itself: A Demo Repo That Refactored Its Own Loop

I pointed my Ralph Loop at its own script to ship a public demo. The useful part was not the recursion — it was every decision I made while the loop made the git log lie, burned $4 on an impossible spec, and taught me which corner I should have cut instead.

shipped

BUILD May 23, 2026 1 of 2

Hardening the Ralph Loop: The 68K-Token Tax in My Overnight Coding Agent

My first overnight Claude Code loop produced real commits and still failed: it burned quota faster than doing the work by hand. The real leak was the 68K-token runtime tax hiding before my prompt.

shipped

BUILD Mar 30, 2026

The Voice Layer

How Metis learned to speak — and why a human had to listen four times before it got the main word right.

shipped

WRITE Mar 28, 2026

ai-agentscareerwritingknowledge-management

7 of 7

The Breed

The AI era didn't create new knowledge. It created an expression layer for people who already had it. This is for that breed.

WRITE Mar 28, 2026

knowledge-managementmetricsworkflow-automation

6 of 7

Three Months of Capture

What the vault actually looks like after three months of PARA and three weeks of automated routing. Real numbers from a real system.

BUILD Mar 28, 2026 5 of 7

The Output Layer

Where the vault speaks back. The pipeline from captured knowledge to published content — and why the blank page problem is gone.

shipped

BUILD Mar 28, 2026 4 of 7

Lyra — The AI Pipe

What the AI in this system actually does — and what it doesn't. Obsidian holds the knowledge. Lyra moves it.

shipped

DECISION Mar 28, 2026 3 of 7

Context: Sixteen years of professional knowledge with no expression layer. Five tools tried — EverNote, OneNote, Notion, Twos, and two years of unstructured Obsidian experiments — none solved the fundamental problem. The knowledge accumulated. The words never came out right.

Why PARA. Why Obsidian.

→ PARA applied to Obsidian on January 6, 2026, for structure. NanoClaw connected on March 9, 2026, for access. Plain Markdown files as the substrate — readable by AI agents without any API, authentication, or rate limits.

Consequence: Capture became zero-friction via WhatsApp. The knowledge base became programmatically readable. The expression problem dissolved: instead of sitting down to write, I describe what happened. The vault routes it. The structure holds it. The AI layer expresses it.

BUILD Mar 28, 2026 2 of 7

The Capture Layer

How a WhatsApp message at 11pm becomes a structured knowledge base entry by morning — the raw input layer of the second brain.

shipped

WRITE Mar 28, 2026

ai-agentsknowledge-managementwriting

1 of 7

The LLM Smell Problem

Someone said my blog has LLM smell. They were right about the writing. They were wrong about where the knowledge came from.

WRITE Mar 26, 2026

ai-workflowimage-generationgeminicloudflare-r2

5 of 5

An AI Pipeline That Reads Your Post Before It Draws the Picture

How to generate blog cover images that actually reflect your content — using a three-tool pipeline built around visual metaphor, not stock photo logic.

WRITE Mar 26, 2026

ai-toolsdesignclaude-skillsux

3 of 5

Five World-Class Designers in Your IDE

Design Council is a Claude Skill that activates five distinct philosophical lenses on demand — Ive, Rams, Hadid, Ingels, Norman — each asking a different question about the same design.

WRITE Mar 26, 2026

ai-workflowdev-toolsclaude-codebmad

4 of 5

The Dev Workflow That Follows You Everywhere

How NanoClaw, the BMAD Method, and Claude Code Remote Control combine into a single workflow that goes from late-night insight to shipped code — regardless of what device you are on.

BUILD Mar 25, 2026 2 of 5

Claude Skills

Open-source Claude Skills for design, architecture, and developer workflows.

shipped

Explore the full archive →

The Orchestration Dividend: What a Month of Mission-Command Cost, to the Cent