DYAI2025/Plumbline

6 stars · Last commit 2026-08-03

Plumbline — a self-learning, customer-value-governed agile AI agent team for Claude Code. 87 subagents + skills, TDD defense-in-depth gates, Kaizen retros, a four-body adversarial council, and an empirically benchmarked QA harness. Does it hang true?

README preview

<div align="center">

# Plumbline

### *Does it hang true?*

**A truth-oriented agent framework for Claude Code that treats green tests as evidence, not as proof of value.**

`86 subagents` · `16 vendored skills` · `11 slash commands` · `/agileteam` governance · `/concilium` council · `Runtime Integrity / Reality Layer` · `benchmarked with explicit evidence ceilings`

Plumbline is built for software work where the expensive failure is not “the test failed.” The expensive failure is: **the test passed, the artifact looked complete, but the real user problem was still not solved.**

[![Plumbline — Plan → Code → Test → ⏸ pause → Deploy: catches the gap before shipping](docs/assets/plumbline-hero-poster.jpg)](docs/assets/plumbline-hero.mp4)

***▶ 10-second overview*** — *click to play. “Tests passed. Value unclear.” → Plumbline catches the gap before shipping.*

**[Live Agent Explorer](https://dyai2025.github.io/Plumbline/)** · inspect the 86 subagents in a browser, no install required.

<br/>

View full repository on GitHub →