About DAF Directory
What This Is
DAF Directory is the AI agent tools directory from Dark Agent Factory. We index every AI coding agent, framework, SDK, and platform we can find — and rate each one on the Dark Factory Score.
The Dark Factory Score is a 0-to-5 evaluation based on Dan Shapiro's “Five Levels of AI-Assisted Programming” framework. Level 0 is manual coding. Level 5 is the dark factory — fully autonomous software production. Every tool lands somewhere on that scale, and we tell you exactly where and why.
This isn't a marketing directory where you pay for placement. Every score is generated by Claude AI using a consistent five-dimension rubric. The evaluation is transparent — every tool gets a detailed rationale explaining its score.
The Dark Factory Score
The score evaluates how close an AI tool gets you to Level 5 autonomous production. It's not a quality rating or a recommendation — it's a capability assessment.
We score five dimensions:
How independently can the tool operate? Can it execute complex, multi-step tasks without constant human guidance?
Does the tool produce production-grade output? Type safety, error handling, test generation, documentation.
How well does it fit into professional workflows? Git, CI/CD, deployment hooks, monitoring.
Can the tool fix its own mistakes? When tests fail, does it iterate?
How much of the development lifecycle does it cover? Full-stack vs single-task.
Our AI pipeline crawls the tool's website, documentation, and GitHub repository. Claude Sonnet evaluates the gathered data against our rubric and produces a scored breakdown with a written rationale. The entire process is automated — no human bias in the scoring.
Tool makers can dispute scores by claiming their listing and providing additional context. We re-evaluate whenever significant updates are released.
Who Built This
DAF Directory is a product of Dark Agent Factory — the AI-native software brand of Woodstock Software LLC, founded by James Williams in Phoenix, Arizona.
Dark Agent Factory is a software production studio where AI agents handle the engineering and humans handle direction and quality. This directory was built using the same Level 3+ production process we evaluate tools against: spec-driven development, Claude Code as the primary implementation tool, and automated testing as the quality gate.
Learn more at darkagentfactory.ai.
Open Data Philosophy
We believe the AI tools landscape needs transparent, unbiased evaluation. That's why every score includes a full breakdown and AI-generated rationale. That's why tool makers can see exactly how they were evaluated. And that's why the directory is free to browse — no paywalls on public information.
Premium features (API access, comparison reports, featured listings) fund the directory's operation and the AI compute costs for evaluating tools. The scores themselves are never influenced by payment.