Hi everyone,
I am developing QASpec, an open-source, spec-driven CLI tool for quality assurance. The objective of the project is to turn QA intent into persistent, versioned artifacts inside the codebase (such as risk analysis, prerequisites, and test cases) instead of relying on transient chat outputs.
The workflow is divided into three phases:
- Analyze: The tool contrasts a Pull Request or a user story against your defined requirements. It does not treat the PR as the source of truth; it assumes the implementation might contain errors, aiming to uncover ambiguities. The approved specs are stored in the repo to serve as context for future PRs in the same area.
- Test Cases: It drafts the structured test cases based on the approved analysis phase. Each case is traced back to a requirement, and a validator automatically checks the coverage.
- Publish: A final validation gate checks the status and uploads the approved test cases to your test management system (currently supporting Qase via an MCP server integration).
The repository is in early development. I am looking for technical feedback regarding this architecture, specifically on the viability and friction of maintaining QA specifications directly as versioned codebase artifacts.
