Skip to content
View calvinnwq's full-sized avatar

Block or report calvinnwq

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
calvinnwq/README.md

Hi, I'm Calvin 👋

AI-native builder & problem solver. Senior SWE @ Atlassian, based in Sydney 🇦🇺, originally from Singapore 🇸🇬.

I build tools for the messy middle of agentic engineering: planning real work, coordinating specialist agents, validating changes, preserving useful evidence, and keeping humans in the loop without slowing everything down.

Blog · X · LinkedIn

What I'm building

  • ShakedownKit - Ultralight backpacking gear management.
  • Agent Swarm - focused swarms of AI agents with one deterministic synthesis back.Many AI agents on the problem. One answer back.
  • Momentum - a TypeScript CLI for durable autonomous repo-work orchestration, local verification, and goal-to-implementation loops.
  • Skill Suitcase - read-only planning and packaging for portable agent skills across agent homes and toolchains.
  • Artshelf - accountable retention for the temporary files, logs, exports, and evidence that agents leave behind.
  • Reasoning Benchmark - small prompts that test whether models actually understand the question, not just the words in it.

Current Focus

I'm spending most of my personal build time on:

  • making AI-agent workflows more inspectable, restartable, and boringly reliable
  • turning coding agents into systems that can plan, execute, verify, review, and clean up real changes
  • building reusable skills, workflows, and tools that work across Codex, Claude, OpenClaw, and local CLI environments
  • writing about AI, engineering, decision-making, health, and adventure at ngxcalvin.com
  • building ShakedownKit for ultralight hikers who want better gear lists, reusable kits, and shakedown-ready pack planning

What I Care About

I like tools that are small, sharp, and honest about their contract.

Good agent tooling should leave a trail: what it planned, what it changed, what passed, what failed, what needs a human, and how to roll back. The interesting work is not just making agents faster. It is making them dependable enough to trust with real workflows.

Outside the terminal, I’m into mountaineering, hiking, surfing, strength training, and the kind of gear nerdery that makes pack weight spreadsheets feel completely reasonable.

Pinned Loading

  1. agent-swarm agent-swarm Public

    Run focused swarms of AI agents and get one deterministic synthesis back.

    TypeScript 7

  2. artshelf artshelf Public

    Tiny CLI for accountable temporary artifact retention.

    TypeScript 1

  3. momentum momentum Public

    Momentum: TypeScript CLI for driving goal-to-implementation loops with local verification artifacts

    TypeScript

  4. reasoning-benchmark reasoning-benchmark Public

    Does the model actually understand the question, or just the words in it?

    Python