rqm Documentation
rqm is a local, deterministic requirements tool written in Rust. Requirements live as YAML files in your repository, next to the code, with stable IDs and UUIDs, kinds, statuses, parents, and links. The tool reads them strictly and reports every problem with a code and a position.
It also connects requirements to tests. You bind tests to a requirement with selectors or annotations. rqm test runs your own test suites and reads their reports. It judges each requirement against a declared policy and records the run. rqm baseline update promotes the evidence from a clean run into a committed baseline file. A requirement appears as verified only when the baseline contains evidence generated for its current text and policy. The label is computed, never written by hand.
Everything is offline and reproducible: no network, no database, and the same inputs give the same hashes and results. rqm shows that the evidence meets the policy you declared. It does not show that the tests are adequate.
Documentation topics
Getting started
Using the command line
- The command line
- rqm init
- rqm create
- rqm rename
- rqm edit
- rqm list
- rqm show
- rqm check
- rqm coverage
- rqm affected
- rqm tests
- rqm trace
- rqm test
- rqm baseline
The file formats
- The configuration file
- The requirement file
- Kinds and fields
- Parents and inheritance
- Identity and hashes
Tests and requirements
- Binding tests to requirements
- Running suites
- How a run is judged
- Reading test reports
- Labels and manual evidence
- The baseline file
- Promotion: updating the baseline
- Running rqm in CI
How rqm works and why
- Why rqm exists
- Principles
- Terms
- States and exit codes
- Decisions and alternatives
- Architecture
- How rqm is tested
- Known limits