AI + SDLC Experiments · Resource Library
Everything we've
learned
Real experiments. Real code. Real results. Our thinking, experiments, and innovations on AI + CI/CD — grab what you need and start shipping.
ResourcesShowing 10 resources
AFK Builds with 100% Green PRs (Chunk Sidecars Inside the Agent Loop)
Five RalphCI Snake game builds with Chunk sidecars in the Review Gate: AFK agent loops, pre-push CI parity, and 100% green pull requests end-to-end.
7 min23 views
View experiment →7 min · 38 views
Team Onboarding Buddy: Claude Code Skills vs the Wiki Maze
We packaged team onboarding as Claude Code plugin skills with MCP-backed checks. The finding is simple: routing plus verification beats another stack of wiki pages.
View →
11 min · 22 views
Hardening RalphCI loops for open source after the February 2026 study
The self-healing CI-aware AI coding loop that fixes CI failures: what we hardened after the February study, why we shipped MIT open source, and how Build Agent, CI Doctor, and Review Gate keep agents honest against real pipelines.
View →
6 min · 23 views
I Gave Claude Seven Personalities and Pointed Them at the SDLC
Claude LiveCaster shipped with one announcer persona. I added six more, covering the full software development lifecycle, and discovered that YAML-driven personas are the real primitive.
View →
6 min · 90 views
Trending
/loop Gave Claude a Heartbeat. I Gave It a Microphone.
I used Claude Code's /loop command to build a live voice announcer that narrates AI model eval races in real time. 16 commits, one overnight session, and a Ken Squier impression that's better than it has any right to be.
View →
10 min · 55 views
HOT
I Built a Server That Rewrites Itself While Running
A Node.js server where visitors vote on features they want, an AI agent writes the code, and the server deploys it to itself. Live. No restart, no build step, no deployment pipeline.
View →
8 min · 196 views
We Let an AI Agent Say 'I Passed.' Was It Actually Good?
10 controlled experiments reveal that 80% of AI agent code fails CI pipelines when there's no feedback loop. Same agent, same task, same model—the only variable was whether it could see what the pipeline saw. That moved the CI pass rate from 20% to 100%.
View →
12 min · 32 views
One Agent vs. a Team: What Benchmark Data Says About Multi-Agent Debugging
Multi-agent AI coding benchmarks and practical implications
View →
10 min · 23 views
Claude Code's Task Tool: From Sequential to Parallel Work
How Claude Code's Task Tool enables parallel processing, transforming sequential operations into multi-agent orchestration for faster development workflows.
View →
12 min · 27 views
Claude Code and Linear: Your Key Partners in AI-Native Development
How Claude Code and Linear form a natural pair for AI-native development workflows, with CircleCI as the execution layer that validates and deploys AI-generated code safely.
View →