Distribute your AI standards with confidence

Publish skills, prompts, and rules once — then distribute them to every team and tool with built-in approval, versioning, and drift detection.

GOVERNANCE MODEL

Four roles that make distribution trustworthy

ADLC Forge maps to how teams actually distribute AI standards — publishers set the bar, consumers pull with confidence.

Product Domain · Publisher

Product Publisher

e.g. Head of Product Ops

Governs org-wide product assets

Product Domain · Consumer

Product Consumer

e.g. Product Manager

Browses catalog, pulls approved assets

Engineering Domain · Publisher

Engineering Publisher

e.g. Staff Engineer

Governs engineering-domain assets

Engineering Domain · Consumer

Engineering Consumer

e.g. Developer

CLI pull, init, CLAUDE.md URI

ASSET TAXONOMY

Seven asset types, purpose-built

Every artifact your team creates for Claude has a type. ADLC Forge governs each one differently.

Skill

A reusable AI capability unit — a CLAUDE.md action that Claude can execute on demand.

Command

A slash command for Claude Code — callable by name during a conversation.

Template

A scaffolding template for new projects or files — consistent starting points.

Prompt

A standalone prompt for a specific task — reusable across teams and products.

Eval Set

A set of test cases for evaluating AI output quality — quality gates for your prompts.

Playbook

A multi-step workflow combining multiple asset types — orchestrated sequences.

Rules File

A .cursorrules or CLAUDE.md configuration — governs how AI behaves in a codebase.

THE EXPERIENCE

Two views, two audiences

Consumers browse and pull from the catalog. Publishers manage distribution health across products.

Browse View — what consumers see
Catalog screenshot

• Assets organized by ADLC phase

• Searchable with version and status visible

• One-click pull into any product

Manage View — what publishers see
Preview Distribution map screenshot

• Deployment health across all products

• Drift detection at a glance

• Version pinning per product scope

SCOPE HIERARCHY

Complementary to Anthropic's architecture

Anthropic defines four scope levels for Claude configuration. ADLC Forge governs which assets are approved at each level.

Enterprise (org-wide)

ADLC Forge manages

Project

ADLC Forge manages

User

ADLC Forge distributes to

Conversation

Assets active here

Skills Rules Prompts Commands

CLI

First-class terminal experience

Engineers pull approved assets directly into their environment. No context-switching, no manual downloads.

# Initialize ADLC Forge in a project
$ adlc-forge init
# Pull approved assets into your project
$ adlc-forge pull --scope project
# Insert a local asset into the catalog
$ adlc-forge import /skills/code-review.md
# Check drift status across products
$ adlc-forge status --all
COMPARISON

ADLC Forge vs. copy-paste distribution

What changes when you stop copy-pasting and start distributing.

Approval workflow

Copy-Paste Distribution

PR review (ad hoc)

ADLC Forge

Structured phase gates

Drift detection

Copy-Paste Distribution

None

ADLC Forge

Automatic per-product

Version pinning

Copy-Paste Distribution

Git tags / branches

ADLC Forge

Semantic per scope level

Onboarding time

Copy-Paste Distribution

Manual repo clone + docs

ADLC Forge

adlc-forge init (one command)

Audit trail

Copy-Paste Distribution

Git log (unstructured)

ADLC Forge

Per-asset event history

PM visibility

Copy-Paste Distribution

Requires Git access

ADLC Forge

Built-in dashboard

Ready to ship your AI standards?

Start with the free tier. No credit card required.