Personal AI agent tools — zero-dependency Python scripts for Claude Code users.
| Repo | What it is |
|---|---|
| checkgrow | Unified quality framework — methodology, failure patterns, decision flow. The canonical source for the CheckGrow thesis. Start here to understand how all the tools connect. |
| Repo | What it does |
|---|---|
| delivery-gate | Stop hook: process monitor + quality gate. Blocks session close when learning capture is stale. |
| self-audit | pip install-able CLI. Four-dimension text consistency audit (regex-based, no LLM). |
| session-cost | Session cost tracker. L0→L3 layered thresholds, auto monthly reports. |
| Repo | What it is |
|---|---|
| dual-pool-review | AI code review using real people's documented principles, not abstract roles. Fixed pool (convergence) + random pool (divergence), cross-orchestrated. Principle-first with confidence levels — never fabricate a quote. |
Each repo is MIT licensed.