Claude Sdlc Harness

Claude Code SDLC Harness

A self-evolving Software Development Life Cycle (SDLC) enforcement system for AI coding agents. Makes Claude plan before coding, test before shipping, and escalate when uncertain. Measures itself getting better over time.

Built on 15+ years of software engineering and founding engineering experience — battle-tested patterns from real production systems, baked into an AI agent that follows tried-and-true software quality practices so you don't have to enforce them manually.
The Reliable lane is the recommended default. Opus 4.6[1m] max + GPT-5.6 Sol xhigh + Fable 5.1 high — field-proven, actively dogfooded.
The Frontier lane (@frontier, not yet published) uses Opus 5.5 xhigh + GPT-5.6 Sol xhigh + Fable 5.1 high and has cycle data from v1.88.0–v1.99.2 but is experimental.
Built for Claude Code. Using OpenAI's Codex CLI instead? Check out codex-sdlc-wizard. Need privacy-first / any-backend (local Ollama, Azure OpenAI, hosted OSS)? See opencode-sdlc-wizard. (Full ecosystem.)
Two lanes. Pick at install time:

Reliable (@latest) — Opus 4.6[1m] max + GPT-5.6 Sol xhigh + Fable 5.1 high.
Field-proven, stable. Use /model claude-opus-4-6[1m] (/effort max).
The bare claude-opus-4-6 pin (without [1m]) gives only 200K.
Frontier (@frontier, not yet published) — Opus 5.5 xhigh + GPT-5.6 Sol xhigh + Fable 5.1 high.
Latest models, less field data. See AI_SETUP_LANES.md
for full details.
Known defect on both lanes: setup does not auto-invoke on a brand-new
project (#698)
— run it manually on first launch: type / plus the skill name.

Install

Requires Claude Code (Anthropic's CLI for Claude). Install it with the native installer — curl -fsSL https://claude.ai/install.sh | bash — which the official setup docs label *Recommended* and which keeps it auto-updating…