10x-stack-assess

Evaluating your tech stack for AI agent-friendliness involves scoring your project's language, framework, build tool, and deployment against four quality gates: typed, convention-based, popular in training data, and well-documented. This process identifies gaps and generates compensation strategies.

The four quality gates for stack assessment are typed, convention-based, popular in training data, and well-documented. Your project's language, framework, and tooling are scored against these gates to determine compatibility with AI agent workflows.

You evaluate your existing project by running the /10x-stack-assess command, which detects project markers and accesses project files. It scores your test runner, package manager, and other stack components against the four quality gates and outputs a structured assessment.

If your tech stack fails a quality gate, you should apply identified compensation strategies. The stack assessment documents these strategies within instruction files to help overcome specific AI agent compatibility gaps in your project components.

Yes, stack assessment requires access to project files and the detection of project markers. This access allows the evaluation to accurately score your specific build tool, CI/CD, and deployment configurations against the four quality gates.

The tech stack assessment documentation is output as a structured markdown file at context/foundation/stack-assessment.md. This file contains stack scores, gap analysis, and compensation strategies for failed quality gates.