bmad-testarch-nfr

Run the NFR evidence audit after implementation evidence exists. It loads thresholds from your tech spec, PRD, or test-design output, gathers test results, metrics, logs, and scan data, then classifies each NFR as PASS, CONCERNS, or FAIL with a release gate recommendation.

The automated audit evaluates four domains: Security, Performance, Reliability, and Maintainability. Thresholds are elicited using the ADR Quality Readiness Checklist's eight categories, and custom NFR categories can be added via configuration.

No. The audit only evaluates existing implementation evidence such as load test reports, Snyk scans, uptime monitoring data, and CI burn-in logs. Use test-design earlier in the lifecycle to plan NFR thresholds and evidence collection.

Undefined thresholds are marked UNKNOWN and the finding is classified as CONCERNS, never PASS. The audit never guesses thresholds, and the aggregation step downgrades any PASS assigned to an UNKNOWN-threshold category.

It produces an nfr-assessment.md report with per-domain findings, a gate YAML snippet for CI integration, quick wins, prioritized remediation actions with owners and effort estimates, and an evidence-gap checklist with deadlines.

Yes. The workflow tracks progress in the output file's YAML frontmatter. Selecting Resume mode loads the saved state, displays a progress dashboard, and routes directly to the next incomplete step without re-executing completed work.