Lab
Notes, reports and releases
Reports with their sources, tools with their code, and unpolished working notes, newest first. This log is all there is.
- L34-N-008 ronsel: E2E Flows You Can Read, Run and CommitNote Published
flows has grown up and changed its name. ronsel turns end-to-end tests into Markdown documents: prose and executable steps in the same file, run from a notebook-style UI, the CLI or your pipeline.
- L34-N-006 Technical guardrails for LLM-assisted code: the full catalogNote Published
A defense-in-depth reference for Node/NestJS teams: nine layers of automated guardrails, from agent hooks and compiler strictness to CI gates and runtime rollback, each one catching what the previous one missed.
- L34-N-007 redact, Rewritten in Go: One Binary, a TUI, and Reproducible ScansNote Published
The new version of redact ships as a single static Go binary with an interactive TUI and a headless CLI. Same two scanners, plus persisted, reproducible results -- and no Python to set up.
- L34-R-001 More code, less confidenceReport Published
What DORA 2025, GitClear, LinearB and Stack Overflow data say about why scrum teams ship worse code — and the habits of the teams that don't.
- L34-N-005 One review, three tiers: a feature delivery workflow that scales its own overheadNote Published
From business story to development without the queue: a swimlane workflow where architecture, contracts and infrastructure are checked once, asynchronously, and infra never sits on the critical path. Block by block, with a worked example and the sources it comes from.
- L34-N-004 redact: Find Sensitive Data in Your Repos Before It Reaches an LLMNote Published
An open-source tool that scans repositories for secrets, credentials, and PII using regex patterns and local LLM analysis, helping you build guardrail lists before sending code to an AI.
- L34-N-003 proxy: An Open-Source Gateway to Secure and Control Your AI UsageNote Published
An OpenAI-compatible reverse proxy that manages credentials, enforces rate limits, tracks token usage, and applies content guardrails for LLM API requests.
- L34-N-002 The 5 Slides Generative AI Theoretical Training CourseNote Published
A concise theoretical training course on generative AI, condensed into 5 key slides. Delivered for a customer in India.
- L34-N-001 flows: Resilient and Declarative E2E Workflows and automationNote Published
Simple YAML files for complex E2E flows testing.