Living Architecture Nodes for VS Code

Keep codebase architecture memory current before humans or AI change the code.

Living Architecture Nodes is a local-first maintenance protocol for keeping module intent, dependencies, drift, regression memory, and AI/dev handoff context alongside the code that changes.

What the extension does

scans the current workspace for Living Architecture Nodes artifacts;
shows architecture-memory health in the VS Code sidebar;
detects missing .node.md companion files;
detects stale/dirty node memory when source files move ahead of their companions;
detects orphan node files;
creates missing node files as generated drafts after explicit confirmation;
opens ARCH.md, NERVE.md, and CHANGELOG.node.md;
exports local JSON and Markdown handoff bundles;
keeps core operation local with no telemetry.

Generated drafts are not verified architecture truth. They must be populated, reviewed, and verified separately.

Free, Pro, and Team

The Marketplace extension remains Free and the core local workflow does not require an account.

The commercial product model uses Free / Pro / Team. Free remains useful and account-optional. Pro is the individual paid tier, and Team is the organization paid tier.

Current paid capability direction includes semantic architectural drift, impact analysis, regression correlation, signed verification receipts, historical comparison, shared policy, and cross-repository analysis.

Paid capabilities and paid prices are not enabled in the current pre-release channel. Until production pricing, billing, entitlement, and customer-term gates are activated, the extension safely operates as Free.

See PRODUCT-TIERS.md for the public tier contract.

Ordinary local scans are not metered by LAN. A paid subscription unlocks capabilities; it is not a bundle of scan credits. Current Free/Pro/Team checks are designed to execute locally, and entitlement refresh is not required for every scan. Remote compute is not production-enabled.

If a deeper check is unavailable because it was not executed, the correct status is:

NOT VERIFIED

That does not mean failed or unsafe.

Restricted Mode / Workspace Trust

Read-only architecture-memory scanning remains available in Restricted Mode.

Operations that modify the workspace require Workspace Trust:

generating .node.md drafts;
writing diagnostic export bundles.

Write and export destinations are confined to the active workspace and reject path traversal or symbolic-link escape.

Commands

Living Architecture Nodes: Scan Workspace
Living Architecture Nodes: Generate Missing Node Files
Living Architecture Nodes: Export AI/Dev Handoff Bundle
Living Architecture Nodes: Show Product Status
Living Architecture Nodes: Open ARCH.md
Living Architecture Nodes: Open NERVE.md
Living Architecture Nodes: Open CHANGELOG.node.md

Required repository artifacts

A project using Living Architecture Nodes should include:

.node.md companion files per significant module/component
ARCH.md
NERVE.md
CHANGELOG.node.md
diagnostic export interface

Typical workflow

Open a workspace.
Open the Living Architecture Nodes sidebar.
Run Scan Workspace.
Review missing, stale, or orphan architecture memory.
In a trusted workspace, create missing node drafts when useful.
Populate and review drafts before treating them as architecture truth.
Export a local handoff bundle when another human or AI system needs diagnostic context.

Output

Diagnostic exports are written locally to .lan-vscode/ by default:

.lan-vscode/living-architecture-diagnostic.json
.lan-vscode/living-architecture-diagnostic.md

The export path must remain inside the trusted workspace.

Privacy

The scanner uses repository structure, relative paths, companion-node presence, and filesystem metadata; it does not read source-file contents for architecture-memory scanning.

Diagnostic handoff exports are local. They contain counts plus relative paths associated with findings, not the full source/node inventory. Absolute workspace paths are removed and secret-shaped metadata is defensively redacted before both JSON and Markdown are written. Core Free operation does not send source code, architecture memory, diagnostics, secrets, repository content, or behavioral telemetry to a remote service.

Future paid entitlement checks are designed to exchange account/product entitlement metadata only—not repository contents.

See PRIVACY.md and SECURITY.md.

Release channel

Living Architecture Nodes has a stable channel and an optional pre-release channel.

The pre-release channel is used to validate new product, security, entitlement, and upgrade behavior before those changes are promoted to stable.

See CHANGELOG.md for current versions and release details.

Support and licensing

Product page: https://altru.dev/living-architecture-nodes
Specification: https://github.com/altrudev/living-architecture-nodes
GitHub Action: https://github.com/altrudev/living-architecture-nodes-action
Issues/support: https://github.com/altrudev/living-architecture-nodes-vscode/issues
Support policy: SUPPORT.md
Security policy: SECURITY.md
License: LICENSE

Living Architecture Nodes™ is a separate product from other Altru.dev research and development systems.

Developed by Altru.dev — Code For Humanity.