Skip to content

Roadmap

The V-Model Extension Pack is actively developed. Here's where we are, where we've been, and where we're going.

Current Release: v0.7.0

The extension provides 17 commands covering four V-Model levels plus cross-cutting safety/quality concerns and the bridge to deterministic, gated implementation:

V-Model Levels

Level Design Command Test Command Traceability
Requirements ↔ Acceptance requirements acceptance Matrix A
System Design ↔ System Test system-design system-test Matrix B
Architecture ↔ Integration Test architecture-design integration-test Matrix C
Module Design ↔ Unit Test module-design unit-test Matrix D

Cross-Cutting Commands

Command Purpose
trace Bidirectional traceability matrix (A + B + C + D + H)
hazard-analysis ISO 14971/26262 FMEA with HAZ-NNN IDs and Matrix H
impact-analysis Dependency graph traversal for change impact assessment
peer-review AI-powered stateless linter with PRF-{ARTIFACT}-NNN findings
test-results JUnit XML + Cobertura XML ingestor for matrix status updates
audit-report Deterministic release audit report with compliance gating

Bridge to Implementation (v0.7.0)

Command Purpose
plan V-Model-enriched plan.md synthesis with pinned-schema validation
tasks TDD-ordered tasks.md with hazard-driven priority elevation and per-task Implements-directive headers
implement 12-step gated pipeline: 8-stage run-v-model-gate.sh → codegen → 4-level test gen → splice → hallucination guard → trace post-hook

Test coverage

  • 455 BATS tests (Bash) · 431 Pester tests (PowerShell)
  • 89 structural evaluations · 53 LLM evaluations · 32 E2E tests (advisory in v0.7.0)
  • Agent definitions for all 17 commands

Shipped Milestones

Version Date Highlights
v0.1.0 2026-02-19 Extension scaffold, requirements, acceptance, trace commands, three-tier ID schema, helper scripts (Bash + PowerShell)
v0.2.0 2026-02-20 system-design, system-test commands, dual-matrix traceability (A + B), golden examples for medical device and automotive ADAS, E2E evaluation harness
v0.3.0 2026-02-21 architecture-design, integration-test commands, triple-matrix (A + B + C), CROSS-CUTTING module tag, consolidated fixture pattern
v0.4.0 2026-02-22 module-design, unit-test commands, four-tier ID schema, quadruple-matrix (A + B + C + D), id-schema-guide documentation
v0.5.0 2026-04-06 hazard-analysis, impact-analysis, peer-review, test-results, audit-report commands, Matrix H, 14 agent definitions, 4× test growth
v0.6.0 2026-04-23 Domain Overlay Architecture (36 overlay files), ID Lifecycle Model (deprecation + suspect cascade), Standards Enrichment (26 standards, Governing Standards sections in all 11 commands), features 002–006 evolved
v0.7.0 2026-05-03 Bridge commands (plan, tasks, implement), 8-stage run-v-model-gate.sh pipeline, validate-artifact-status (MF-6) and validate-domain-profile (MF-7) validators, splicer hardening (MF-5: id-match, duplicate detection, --region-from, diff -u audit trail), hallucination guard --canonical/--scan/--changed-only, Compliance & Hybrid Modes documentation, 4 lifecycle hooks

For detailed release notes, see the Changelog.

What's Next

v0.8.0 Stabilization Backlog

The pre-v0.8.0 architectural audit identified ten Epics that did not fit the v0.7.x patch envelope. They are tracked in .tmp/pre-v0.8.0-audit/ultimate-stabilization-blueprint.md and summarised in the v0.7.0 release notes:

  • EPIC-1compliance_mode: strict profile (flips hook optional: truefalse; makes hybrid mode mechanically detectable in CI).
  • EPIC-2 — Sealed baseline + signed run attestation (append-only ledger, .specify/v-model/baseline.lock, validate-task-provenance.sh).
  • EPIC-3 — Immutable compliance-artifact protection (snapshot diff in before_implement / after_implement).
  • EPIC-4 — Semantic directive validation (path/symbol/reachability checks layered onto the existence-only check shipped in v0.7.0).
  • EPIC-5 — Single v-model verify aggregator with SARIF/JSON output.
  • EPIC-6 — Authorized managed regions (region IDs encode authorized V IDs; splicer cross-validates against module-design.md's Target Source File(s)).
  • EPIC-7 — Deterministic golden-output e2e fixtures (LLM-judge demoted to advisory regression detector).
  • EPIC-8 — Package-install smoke test (assert all 17 commands + 4 hooks register after a clean install).
  • EPIC-9 — Strict prompt-injection containment (net-new linter for prompt-injection patterns in V-Model artifacts).
  • EPIC-10 — Reviewer/signoff schema (per-artifact reviewers: and signoff: fields validated against domain-specific minimums).

Pre-built Regulatory Template Packs

Domain-specific templates for IEC 62304, ISO 26262, and DO-178C that pre-populate required sections, terminology, and compliance language. Instead of starting from generic templates, teams get standard-specific boilerplate with the correct ASIL/SIL/Class verbiage pre-filled.

Bidirectional ALM Synchronization

Two-way sync with enterprise ALM platforms (Jama Connect, IBM DOORS, Siemens Polarion), eliminating the risk of fragmented sources of truth between Git-based artifacts and the enterprise system of record.

Trend Tracking

Monitor requirement quality scores, coverage percentages, and traceability completeness over time to catch degradation before it becomes an audit finding.

Future Considerations

Under exploration

These items are informed by the architecture and community feedback but are not yet committed to the roadmap:

  • IDE Integration — Editor extensions for VS Code and JetBrains that surface traceability data, coverage status, and peer review findings directly in the development environment
  • Additional Domain Standards — Support for standards beyond IEC 62304, ISO 26262, and DO-178C — such as IEC 61508 (industrial), EN 50128 (railway), and ECSS (space)
  • Visual Dashboard — A web-based dashboard for visualizing traceability matrices, coverage trends, impact analysis graphs, and audit readiness status
  • Report Customization — Assessor-specific formatting and full customization of the audit report output for different regulatory bodies

Influence the Roadmap

Your input shapes what we build next:

  • Feature requestsOpen an issue describing your use case and how it fits the V-Model workflow
  • Discussions — Join the conversation in GitHub Discussions to share ideas, vote on proposals, and connect with other users
  • Contributions — See the Contributing Guide to get started with code, tests, or documentation

The philosophy

The AI drafts. The human decides. The scripts verify. Git remembers.