ALDC — AL Agentic Engineering System

Skills-based, spec-driven, TDD-orchestrated development framework for Microsoft Dynamics 365 Business Central.

Engineering systems, visibly reasoned.

ALDC combines specialist agents, reusable domain skills and human approval to support Business Central extension development from requirements to review.

ALDC 5.2.0 brings improvements to Copilot Chat, Copilot CLI and Codex. It adds explicit security review, BCQuality article dispositions, stronger planning and delegation contracts, and configurable Conductor autonomy and test execution. The Marketplace badge reports the currently published extension version. See what changed and the changelog.

Channel	Current delivery	Scope
VS Code Marketplace	See Marketplace badge	Copilot Chat extension and packaged toolkit; existing projects adopt new contracts through Update Toolkit.
Latest GitHub release	v5.0.3	Tagged source and toolkit archive; a separate delivery from the VSIX.
5.2.0	Root package, Copilot catalogs and generated CLI/Codex distributions 5.2.0	Prepared source; publication to Marketplace, GitHub releases and npm is separate.
Claude Code edition	5.1.1	Maintained independently, with its own agents, hooks, contracts and release evolution.

Claude Code remains part of ALDC, with a shared engineering foundation and its own maintained agents, hooks, contracts and release evolution. Its surface changes are documented separately in the changelog and claude-plugin/SURFACE.md. A version in one channel does not establish an equivalent update in another, or publication to npm.

For an installed Claude Code plugin, refresh the official source and plugin:

claude plugin marketplace update aldc-marketplace
claude plugin update aldc@aldc-marketplace

Open a new Claude Code session and check the version against its plugin catalog. This refreshes the plugin's contracts; project-local toolkit copies need their own reviewed update. See installation and recovery for each surface.

Start here

Choose the installation guide for your host, then follow the Quick Start. Use your own AL project and BC sandbox; check its app.json, symbols and permissions before any build or publication. ALDC supplies development context and workflows, not a Business Central environment.

Start with the toolkit installed in your project and a bounded requirement. Architect and Spec prepare the design and technical specification for your review before implementation.

Architecture and distribution map · Installation/recovery · Doctor · Public BC/AL/MCP repository catalogue.

What changed in 5.2.0

Reviews now include five mandatory security checks and a disposition for every selected BCQuality article: apply it or explain why it does not apply. Agent SDK reviews check completeness when the app extends Agent Metadata Provider. Architect, Spec Agent and Planning Subagent cite the knowledge informing their decisions. Planning records the standard publisher flow and the chosen event; subscribers explain the choice above their attribute.

Conductor delegates research, each implementation phase and review to their owners. It records a Mini-design for eligible LOW work without an architecture document; the Core v1.2 architecture gate for MEDIUM/HIGH remains. Final reports distinguish verification performed, checks pending, decisions and open risks.

aldc.yaml adds autonomy.gates: human | pre-approved and testing.runner. Human is the default. With the requester's prior approval, gates become recorded checkpoints in decisions.md; hard stops still produce blocked.md. autonomy.commit remains false by default. A configured test runner supplies actual per-test results after compilation; without one, execution is not-run. ALDC_GATES and ALDC_TEST_RUNNER can override the corresponding settings where the host exposes environment variables. These contracts add no host hooks.

The companion VS Code Project Manager offers an explicit BCQuality mode selector for existing projects and native VS Code installation help. Fresh extension installations select BCQuality plugin mode. The canonical aldc.yaml keeps its external-multiroot default. Project Manager's Human/default and Automatic options map to autonomy.gates: human and pre-approved; Automatic requires prior approval and grants no additional commit or publication permission.

Fresh installations in both Chat profiles select BCQuality plugin mode with automatic discovery and no fixed plugin version pin. Updates preserve the project's selected mode, availability preference and provider identity; customized configuration follows the existing collision policy. A newly seeded workspace includes an external knowledge folder only for an enabled external mode and uses its configured path. Existing workspace files remain owned by the project.

Both Chat profiles retain 5.0.4's dependency guidance: resolve only libraries used by the planned tests from approved dependencies, declared versions and symbols. Claude Code stays on its independently maintained 5.1.1 edition.

Use Project Manager in VS Code

1. Update the extension to 5.2.0 and run AL Collection: Open Project Manager.
2. Select your solution, preview its toolkit installation or update, review customizations and confirm. Updating the extension does not replace the toolkit already installed in a project.
3. In BCQuality, choose Plugin — recommended, External knowledge folder or Disabled — native review, then select Change mode. Install BCQuality… copies the official repository URL for the native source prompt; paste it there. Hosts without that command open the installation guide. Manage plugins opens the host's inventory.
4. In Workflow autonomy, choose Human — approve each checkpoint or Automatic — prior approval, then select Apply mode. Start a new chat with AL Development Conductor.

Human waits at approval checkpoints. Automatic records those decisions and continues within the approved task, retaining hard stops. Changes preserve commit and publication permissions and keep a recoverable configuration copy. A retained older Conductor must be updated before using the autonomy control. Plugin installation and loading are verified by the host.

These Project Manager controls apply to VS Code/Copilot Chat. The other hosts keep their own configuration and permissions. Canvas remains a separate experimental companion.

Earlier Copilot Chat improvements

5.0.3 — knowledge before implementation. Architect and the independent Spec Agent consult BCQuality through al-knowledge, read complete articles and record the evidence used in their decisions. Hosts without a skill invocation tool follow the instructions inline. BCQuality is installed separately; unavailable knowledge and manual fallbacks remain explicit. Windows validation and generated-path handling also improve.

5.0.4 — native AL tools for your project. Copilot Chat's bc29-native Developer and Implementation Subagent can query object IDs within approved project ranges. IDs are not reserved, and the local scan excludes installed extensions. Architect/Spec research stays bounded to relevant symbols and dependencies; native diagnostics use the compatible AL18.1 selector. Developer can compile the selected AL project and consult its diagnostics as part of the implementation workflow.

Conductor keeps coordination and Reviewer keeps query permissions. Publication requires explicit authorization. Graph, LSP, profiling and snapshot operations remain optional and depend on the installed host and tools. The native tooling guide explains the available capabilities, project selection and permissions. Claude Code's later surface releases follow their own contracts.

To update an existing VS Code project, update the extension, run AL Collection: Update Toolkit, review the preview and managed-file collisions, then run AL Collection: Verify Installed Toolkit and reload the window. Each terminal plugin has its own update route. Earlier improvements, including Copilot Chat's Sonnet 5.5 selection, remain documented under their original changelog versions.

Capability	What it provides
Architect and Spec Agent	Approved design, bounded Specs, explicit dependencies and joint consistency review before implementation.
Conductor	Planning, implementation and review coordinated through existing human gates.
Doctor	Read-only diagnosis of configuration and operation-specific runtime observations.
Recommended BCQuality	Plugin or external multiroot reviews, with discovery, loading, execution and index evidence distinguished.
BC28 / BC29-native	Explicit Chat profile selection with role-specific tools and AL18 guidance.
Host adapters	Dedicated content for Copilot Chat, Copilot CLI, Claude Code and Codex.
Recoverable installation	Preview, visible collisions, receipts, verification and restoration.

See Spec Agent, profiles, plugins and recovery and Doctor.

---

Why ALDC?

AI accelerates AL development — but raw code generation is unpredictable, hard to review, and easy to get wrong against Business Central's extension rules. ALDC adds the engineering discipline around it.

Without ALDC	With ALDC
Ad-hoc, one-shot code generation	Contract-driven development (architecture → spec → human approval → test-plan → code)
No checkpoints — you find out at the end	Human-in-the-loop gates at every phase
Tests written last (or never)	TDD-orchestrated — tests FIRST, then code
Base-app edits sneak in	Extension-only by construction
"Looks fine to me" reviews	Cited reviews & audits against BC knowledge

---

What is ALDC?

ALDC (AL Development Collection) transforms how you develop Business Central extensions. Instead of ad-hoc code generation, it provides structured, contract-driven development with specialized agents, composable skills, and human-in-the-loop gates — with host-specific distributions:

• GitHub Copilot — Agents, skills, prompts, and instructions in .github/ and agents/
• Claude Code — Agents, skills, rules, and hooks in .claude/ and claude-plugin/
• Copilot CLI — Terminal agents and commands in copilot-cli-plugin/
• Codex — Generated agent profiles and a local ALDC skill in plugins/aldc-codex/

---

Updating ALDC

The VS Code extension and each host plugin are separate distribution channels. Updating the VSIX does not refresh installed Claude Code, Copilot CLI or Codex plugins.

In VS Code, open AL Collection: Open Project Manager (also available as a view in the Explorer) to preview and confirm the update, verify the installation, restore the previous transaction and run Doctor without leaving the panel; the ALDC Visor view lists the plan artifacts of each requirement with template-specific icons. The palette command AL Collection: Update Toolkit remains available: review file collisions, verify the installation and reload the window. Existing installations retain their recorded profile. The extension's optional al-collection.autoInstall setting is false by default; when enabled, it attempts installation on AL project detection. Toolkit restoration covers installation files, not Business Central data or all project history. See installation and recovery.

Installation

Choose your surface — installation guide · Guía en castellano

For the opt-in BC29 / AL18 native profile in Copilot Chat, see installation, role changes and local validation. BC28 remains the installer default. Claude Code and the dedicated aldc-cli plugin include terminal-specific BC29/AL18 contracts; see the same guide for checkout-local installation and pending runtime verification. The VS Code extension is published separately.

GitHub Copilot

Install from the VS Code Marketplace or:

code --install-extension javierarmestogonzalez.al-development-collection

Then, from the Command Palette:

• AL Collection: Install Toolkit to Workspace — copies the framework into your project's .github/
• AL Collection: Update Toolkit — updates toolkit content; review collision behavior in your installed version
• AL Collection: Validate Installation — verifies compliance

GitHub Copilot (Plugin Marketplace)

ALDC also ships as an installable Copilot plugin — plugin.json declares the agents (agents/), skills (skills/), and prompts (prompts/) that get installed, for editors that support the Copilot plugin marketplace:

The root plugin retains the VS Code-oriented primitives. For Copilot CLI, use the dedicated terminal distribution from this checkout:

copilot plugin install ./copilot-cli-plugin
copilot plugin list

The catalog entry is aldc-cli@aldc-marketplace. See plugin loading checks and limitations before testing. The directory is generated with npm run sync:copilot-cli.

Claude Code (Plugin)

/plugin install aldc
/aldc:al-initialize

al-initialize copies path-scoped rules to .claude/rules/, generates a project CLAUDE.md, and configures the workspace.

Claude Code (Direct)

Clone this repo and open it with Claude Code. The .claude/ directory and CLAUDE.md are detected automatically.

---

Quick Start

GitHub Copilot

1. Install the extension and open your AL project
2. Run AL Collection: Install Toolkit to Workspace
3. Start with @workspace use al-spec.create plus your requirement
4. Follow the guided flow

Claude Code

1. /plugin install aldc
2. /aldc:al-initialize
3. Call any agent: @al-architect, @al-developer, @al-conductor, @al-presales
4. Or run a workflow: /aldc:al-spec-create

See QUICKSTART.md for the full onboarding guide.

---

Key Features

🤖 5 Core Agents — one specialist per development phase

Agent	Role
@AL Architecture & Design Specialist	Designs solutions, information flows, technical decisions
@AL Spec Agent	Specifies approved designs through the shared al-spec.create contract
@AL Implementation Specialist	Implements, debugs, quick adjustments
@AL Development Conductor	Orchestrates TDD implementation with subagents
@AL Pre-Sales & Project Estimation Specialist	Estimation and scoping

🔬 3 Internal Subagents — autonomous specialists inside the Conductor

• AL Planning Subagent — research and context gathering
• AL Implementation Subagent — TDD-only (tests FIRST, code SECOND)
• AL Code Review Subagent — review against spec + architecture

🩺 3 On-demand Specialists — user-invocable, read-only on code

• @AL Triage — reactive diagnosis: reproduce → root-cause → minimal-fix recommendation
• @Dredd — independent auditor: BCQuality-cited static audit with an advisory verdict
• @AL Developer Reviewer — independent review of direct Developer increments

🧠 Core domain skills — knowledge loaded on demand

• Required: api · copilot · debug · performance · events · permissions · testing
• Recommended: migrate · pages · translate · estimation
• The complete source has 16 skill directories, including contribution assistance, BC Agent modules and the shared AL review pipeline. See skills/. Host distributions also package role entries and workflows; their folder counts measure a different scope.

⚙️ 6 Workflows — automated processes

al-spec.create · al-build · al-pr-prepare · al-context.create · al-memory.create · al-initialize

📐 Scoped instructions and the Copilot entrypoint

Eight *.instructions.md files apply by their declared globs, including the optional BC Agent toolkit rule. copilot-instructions.md is the always-on entrypoint and index.md is documentation; neither is an additional scoped rule. See instructions/.

📚 BCQuality (recommended) — external, citable BC knowledge layer

Optional cited review through an explicitly selected host plugin or external multiroot knowledge base. Agents distinguish discovery, loading, execution and best-effort index generation; unavailable providers retain native A–G coverage. See docs/bcquality.md.

📄 Contracts per Requirement — structured docs in .github/plans/{req_name}/

• {req_name}.architecture.md — solution design (from architect)
• {req_name}.spec.md — technical blueprint (from spec.create)
• {req_name}.test-plan.md — test strategy
• memory.md — global context across sessions

---

How It Works

Development Flow

LOW complexity:
  al-spec.create → @AL Implementation Specialist

MEDIUM/HIGH complexity:
  @AL Architecture & Design Specialist → al-spec.create → @AL Development Conductor

The architect designs the solution and can decompose complex requirements into multiple specs, each implemented independently by the conductor.

flowchart TD
    REQ[Requirement] --> CLASSIFY{Complexity?}
    CLASSIFY -->|LOW| SPEC_LOW[al-spec.create]
    SPEC_LOW --> DEV["@AL Implementation Specialist"]

    CLASSIFY -->|MEDIUM/HIGH| ARCH["@AL Architecture & Design Specialist"]
    ARCH -->|Designs solution| ARCH_DOC[architecture.md]
    ARCH --> DECOMPOSE{Decompose?}
    DECOMPOSE -->|Yes| SPEC_A[al-spec.create → spec-A]
    DECOMPOSE -->|Yes| SPEC_B[al-spec.create → spec-B]
    DECOMPOSE -->|No| SPEC_SINGLE[al-spec.create → spec.md]
    SPEC_A --> COND_A["@AL Development Conductor"]
    SPEC_B --> COND_B["@AL Development Conductor"]
    SPEC_SINGLE --> COND["@AL Development Conductor"]

TDD Orchestration

The conductor enforces Test-Driven Development:

flowchart LR
    PLAN[Phase 1\nPlanning] --> IMPL[Phase N\nImplementation]
    IMPL --> REVIEW[Review\nSubagent]
    REVIEW --> HITL{HITL Gate}
    HITL -->|Approved| NEXT[Next Phase]
    HITL -->|Revise| IMPL

    subgraph IMPL_DETAIL[Implement Subagent - TDD]
        RED[RED\nWrite failing tests] --> GREEN[GREEN\nMinimal code]
        GREEN --> REFACTOR[REFACTOR\nAL patterns]
    end

1. Planning subagent researches context
2. Implement subagent creates tests FIRST (RED)
3. Implement subagent writes code to pass tests (GREEN)
4. Implement subagent refactors to AL patterns (REFACTOR)
5. Review subagent validates against spec + architecture
6. Human approves each phase (HITL gate)

Framework Architecture

graph TB
    subgraph PUBLIC["Public Agents (user-invocable)"]
        ARCH["@AL Architecture & Design Specialist"]
        DEV["@AL Implementation Specialist"]
        COND["@AL Development Conductor"]
        PRE["@AL Pre-Sales & Project Estimation Specialist"]
    end

    subgraph INTERNAL["Internal Subagents (conductor-only)"]
        PLAN[AL Planning Subagent]
        IMPL[AL Implementation Subagent]
        REV[AL Code Review Subagent]
    end

    subgraph SKILLS["11 Composable Skills"]
        S1[skill-api]
        S2[skill-copilot]
        S3[skill-debug]
        S4[skill-performance]
        S5[skill-events]
        S6[skill-permissions]
        S7[skill-testing]
        S8[skill-migrate]
        S9[skill-pages]
        S10[skill-translate]
        S11[skill-estimation]
    end

    subgraph WORKFLOWS["6 Retained Workflows"]
        W1[al-spec.create]
        W2[al-build]
        W3[al-pr-prepare]
        W4[al-context.create]
        W5[al-memory.create]
        W6[al-initialize]
    end

    ARCH --> SKILLS
    DEV --> SKILLS
    COND --> INTERNAL
    PRE --> SKILLS

    W1 --> ARCH
    W1 --> COND

Contract Structure

.github/
└── plans/
    ├── memory.md                          ← Global (cross-session context)
    └── {req_name}/
        ├── {req_name}.architecture.md    ← From @AL Architecture & Design Specialist
        ├── {req_name}.spec.md            ← From al-spec.create
        ├── {req_name}.test-plan.md       ← From al-spec.create or conductor
        ├── {req_name}-plan.md            ← From @AL Development Conductor (Planning)
        ├── {req_name}-phase-1-complete.md
        └── {req_name}-phase-N-complete.md

---

Routing Guide

Complexity	Route	When
LOW	al-spec.create → @AL Implementation Specialist	Simple field, validation, single UI change
MEDIUM	@AL Architecture & Design Specialist → al-spec.create → @AL Development Conductor	Business logic, event-driven feature
HIGH	@AL Architecture & Design Specialist → al-spec.create → @AL Development Conductor	Multi-module, external integration, architectural change

Not sure where to start?

@AL Architecture & Design Specialist

I need to [describe your requirement]

The architect analyzes requirements, designs the solution, and recommends the appropriate workflow.

---

ALDC for Claude Code

ALDC is available as a native Claude Code integration in two forms:

• Official Plugin (claude-plugin/) — Install with /plugin install aldc, namespaced as aldc:*
• Direct Integration (.claude/) — Auto-detected when opening the repo in Claude Code

What's Included

Primitive	Direct (.claude/)	Plugin (aldc:)	Count
Agents	.claude/agents/	agents/	9 public + 3 internal
Skills	.claude/skills/	skills/	16 knowledge modules + 9 role entries + 11 workflows
Rules	.claude/rules/	rules/ (injected via al-initialize)	8 coding standards
MCP Servers	.mcp.json	.mcp.json	3 servers
Hooks	.claude/settings.json	hooks/hooks.json	2 hooks
Instructions	CLAUDE.md	CLAUDE.md	Agent routing, workflows

How It Maps

GitHub Copilot              →  Claude Code (Direct)         →  Claude Code (Plugin)
──────────────────────────────────────────────────────────────────────────────────────
agents/*.agent.md           →  .claude/agents/*.md          →  agents/*.md
skills/*/SKILL.md           →  .claude/skills/*/SKILL.md    →  skills/*/SKILL.md
instructions/*.md           →  .claude/rules/*.md           →  rules/*.md
prompts/*.prompt.md         →  .claude/skills/ (workflows)  →  skills/ (workflows)
.github/copilot-instructions.md → CLAUDE.md                 →  plugin.json + CLAUDE.md

Agent Routing (Claude Code)

Agent	Direct	Plugin
Architecture & Design	@al-architect	@aldc:al-architect
Technical Specification	@al-spec-agent	@aldc:al-spec-agent
Implementation	@al-developer	@aldc:al-developer
TDD Orchestration	@al-conductor	@aldc:al-conductor
Estimation & Scoping	@al-presales	@aldc:al-presales
Agent Builder	@al-agent-builder	@aldc:al-agent-builder

Workflows (Claude Code)

Workflow	Direct	Plugin
Create specifications	/al-spec-create	/aldc:al-spec-create
Build & deploy	/al-build	/aldc:al-build
Prepare PR	/al-pr-prepare	/aldc:al-pr-prepare
Session memory	/al-memory-create	/aldc:al-memory-create
Project context	/al-context-create	/aldc:al-context-create
Environment setup	/al-initialize	/aldc:al-initialize

Hooks

Claude Code hooks enforce quality gates automatically:

• PostToolUse (Write/Edit) — Reminds to run tests after file modifications
• Stop — Reminds to verify Skills Evidencing was declared

Plugin User Configuration

On first enable, the plugin prompts for optional settings:

Setting	Description
bcSandboxUrl	URL of your Business Central sandbox environment
publisherName	Your extension publisher name for app.json

---

Using BCQuality (recommended)

BCQuality supports explicit plugin and external-multiroot modes in aldc.yaml. Plugin mode loads the configured skill (default al-code-review); multiroot retains the external Entry workflow. An expected version or commit is not proof of the installed identity, and catalog discovery is not execution. When unavailable, native review continues. See configuration and evidence.

BC Agent Builder (optional)

Build Business Central Agents with the AI Development Toolkit and Agent SDK. Includes: @AL Agent Builder agent, 3 skills, 4 workflows, validation tools. See BC Agent Builder documentation.

---

Validation

node tools/aldc-validate/index.js --config aldc.yaml

Expected result: ✅ ALDC Core v1.2 COMPLIANT

---

File Structure

AL-Development-Collection-for-GitHub-Copilot/
│
│── GitHub Copilot ─────────────────────────────────────
├── .github/
│   ├── copilot-instructions.md           # Master coordination
│   └── plans/                            # Per-requirement contracts
│       ├── memory.md                     # Global memory (cross-session)
│       └── {req_name}/
│           ├── {req_name}.architecture.md
│           ├── {req_name}.spec.md
│           └── {req_name}.test-plan.md
├── agents/                               # 12 agents (5 core + 3 on-demand + 3 subagents + 1 extension)
├── skills/                               # 16 composable skill directories
├── prompts/                              # 11 workflows (6 core + 5 BC Agents pack)
├── instructions/                         # 9 scoped rules + entrypoint (10 contracts); documentation excluded
│
│── Claude Code (Direct) ───────────────────────────────
├── CLAUDE.md                             # Master instructions
├── .mcp.json                             # MCP server configuration
├── .claude/
│   ├── agents/                           # 12 agents (9 public + 3 internal)
│   ├── skills/                           # 16 skills (composable knowledge modules)
│   ├── rules/                            # 8 path-scoped coding standards
│   └── settings.json                     # Hooks + permissions
│
│── Claude Code Plugin ─────────────────────────────────
├── claude-plugin/                        # independently maintained Claude Code edition
│   ├── .claude-plugin/plugin.json        # Plugin manifest
│   ├── agents/                           # 12 agents (auto-discovered)
│   ├── skills/                           # 9 role entries + 11 workflows + 16 knowledge
│   ├── hooks/hooks.json                  # PostToolUse + Stop hooks
│   ├── rules/                            # 8 rules (injected via al-initialize)
│   ├── .mcp.json                         # 3 MCP servers
│   └── README.md                         # Plugin documentation
│
│── Shared ─────────────────────────────────────────────
├── docs/
│   ├── framework/                        # Normative spec + diagrams
│   └── templates/                        # Immutable contract templates (7)
├── tools/aldc-validate/                  # ALDC Core validator
├── aldc.yaml                             # Core v1.2 configuration
├── CHANGELOG.md                          # Version history
└── README.md                             # This file

---

Reference

Framework Documentation

• Core Specification v1.2
• Architecture Diagrams
• Manifesto
• Quickstart
• Governance
• Compliance Model
• Migration Guide v1.0→v1.1

MCP Servers Integration

Server	Purpose
al-symbols-mcp	AL object analysis from compiled .app packages
context7	Up-to-date library documentation retrieval
microsoft-docs	Official Microsoft/Azure documentation search

Requirements

GitHub Copilot

• Visual Studio Code 1.85.0+
• GitHub Copilot (agent and skill features)
• AL Language Extension
• Node.js 20+ (for validator)

Claude Code

• Claude Code CLI v1.0.33+
• AL Language Extension
• Node.js 20+ (for MCP servers via npx)

---

What's New

5.0.0 — one delivery across every surface

Dedicated Spec Agent, per-operation Doctor, BC29-native tooling, terminal adapters and recoverable initialization, plus the VS Code Project Manager and Visor, BCQuality read as context in the design phases, and a Claude Code plugin that is generated rather than hand-written. It supersedes the unreleased 4.3.0 and 4.3.1, neither of which was ever tagged or published. Major for the plugin layout and the per-surface plans and audits roots; see the 5.0.0 changelog.

4.2.0 — Conformance release

The framework now enforces its own spec in CI.

• Core Spec v1.2 (original 4.2.0 release) — originally normalized the tier model to 4 core agents + 2 on-demand (al-triage, dredd) + 3 subagents + 1 extension (al-agent-builder); 16 skills; 11 workflows. The subsequent canonical Spec Agent increment adds a fifth core role (11 total); Developer Reviewer adds a direct review role (12 total, with a shared review pipeline).
• Conformance tooling — scripts/check-conformance.js (counters, cross-references, links, frontmatter) and scripts/sync-foundation.js --check (zero drift between the canonical trees and packages/foundation/) run on every push and PR.
• ARCHITECTURE.md — one-page map of what is source, what is generated, and which distribution channel consumes each tree.
• Fixed: truncated skill-manifest in packages/foundation/, broken README links, undeclared primitives in aldc.yaml, contradictory counters.

4.1.0 — Lower token cost & cited audits

• ⚡ Lower token / AIC cost — trimmed always-on entrypoint (~31% lighter), narrow instruction globs (applyTo by object type), curated context passing, condensed primitives, BCQuality task-context built once and passed inline.
• 📚 Cited reviews & audits with BCQuality (recommended) — agents back findings with a pinned BC knowledge base; graceful native fallback (never blocks).
• @AL Triage and @Dredd — read-only on-demand specialists.
• skill-contribution-assistant — guided contribution workflow.
• Restored full architecture & spec templates with authoring guidance.

4.0.0 — Token efficiency & foundation layout

• Agents, instructions, skills, prompts, and templates condensed for a lower token footprint — behavior preserved.
• New packages/foundation/ layout; Architecture Decision Records (docs/decisions/); new al-agent.build-instructions prompt; English-only content.
• Breaking: primitives also available under packages/foundation/; agent/instruction/skill wording token-optimized (behavior unchanged).

See CHANGELOG.md for full details.

---

Author

Javier Armesto González Microsoft MVP (Business Central & Azure AI Services) Head of R&D & AI at VS Sistemas LinkedIn · Tech Sphere Dynamics

---

Support & Contributing

• Report issues: GitHub Issues
• Ask questions: GitHub Discussions
• See CONTRIBUTING.md for contribution guidelines

---

License

MIT — See LICENSE for details.

Status: ALDC Core v1.2 COMPLIANT · Distributions: Copilot Chat / CLI, Claude Code, Codex · Root package: 5.2.0 · Claude Code: 5.1.1 · VS Code Marketplace: see badge

Independent review

Use Developer Reviewer after a direct Developer increment; use the Conductor's review subagent for orchestrated phases. Dredd remains the independent advisory auditor for an explicit file set, changes or a broader codebase. All three share provider execution and coverage rules: loading BCQuality is not execution, and skipping its optional index refresh does not stop path-based review. A partial review never becomes approval from zero findings. See the review contract.