Skip to content

RIGHT_SIDE_OF_V.md

Latest commit

 

History

History
48 lines (38 loc) · 2.23 KB

RIGHT_SIDE_OF_V.md

File metadata and controls

48 lines (38 loc) · 2.23 KB

Right Side of the V: Verification & Validation

This repo already provides strong “left side of the V” support via:

  • disciplined-research (Phase 1)
  • disciplined-design (Phase 2)
  • disciplined-specification (Phase 2.5)
  • disciplined-implementation (Phase 3)

It also includes specialist verification skills (code-review, security-audit, rust-performance, testing). The gap was an explicit, disciplined right-side-of-V workflow that:

  1. selects the right gates based on risk,
  2. ties evidence back to requirements, and
  3. covers UAT and visual regression explicitly.

Proposed Agent/Skill Design

Treat “agents” as roles, implemented as skills.

Agent Role Skill Primary Output
V&V Lead / Quality Gate quality-gate Quality Gate Report (go/no-go + evidence)
Static Analysis Engineer ubs-scanner Automated bug detection (1000+ patterns)
Code Reviewer code-review Review findings + checklist verification
Security Reviewer security-audit Security findings + remediation
Performance Reviewer rust-performance Benchmarks/profiles + regression risk
Traceability Engineer requirements-traceability Traceability matrix + coverage gaps
UAT Engineer acceptance-testing UAT plan + acceptance scenarios + sign-off
Visual QA Engineer visual-testing Visual regression plan/tests + baseline policy

Disciplined Right-Side Flow (After Implementation)

  1. Run quality-gate on the PR/change.
  2. quality-gate selects which specialist gates apply (always: review + traceability; conditional: security/perf/UAT/visual).
  3. Produce a single Quality Gate Report that links:
    • Requirement IDs → implementation locations → tests → evidence artifacts
  4. Only merge/release when blockers are resolved or explicitly waived with rationale.

Evidence Pack (Recommended)

For any non-trivial change, the goal is to be able to answer:

  • “Which requirements did we change?”
  • “Where is it implemented?”
  • “How do we know it works?”
  • “How do we know it’s safe and non-regressing?”

The skills are designed to produce that evidence in a compact, reviewable form (usually Markdown reports plus logs/screenshots/bench results).