Deliver code with ADLC5
Public preview. Workflow contracts are tested; comparative coding-quality results and outside-contributor qualification remain pending.
ADLC5 gives AI coding tools a repeatable delivery workflow: Specify → Plan → Tasks → Implement, connected by a fifth pillar: Intelligence. Intelligence gathers repository knowledge, assembles focused context, and grounds progression in current evidence. Agent skills handle judgment; one deterministic kernel owns progression and checks. It aims for repeatable acceptance and failure handling, not identical generated code.
Read why an AI coding tool needs a delivery workflow.
Use it in your application repository. This repository distributes the framework
and a portable dogfood/ consumer example.
Five pillars, connected delivery
Specify · Plan · Tasks · Implement · Intelligence. The first four are delivery stages; Intelligence connects them to your codebase. Models supply reasoning; ADLC5 supplies structured knowledge, focused context, and deterministic checks.
| Intelligence capability | Role in delivery |
|---|---|
| Repository maps | Locate relevant modules, symbols, import dependencies, and tests before planning. |
| Reviewed guidance | Carry architecture, boundaries, and commands across agent sessions. |
| Focused context packs | Equip each agent with its story, acceptance, and repository guidance. |
| Freshness and evidence | Refresh stale maps and require current checks and review before completion. |
The distinction is how these capabilities connect the lifecycle. Read Intelligence for a worked flow, commands, and honest limits. SOUL supports this pillar as the reasoning discipline across the four stages.
Start here
- Install ADLC5 for your agent host.
- Initialize your workspace and feature.
- Invoke
@adlc5 for my-featureand follow the four-stage lifecycle.
Proportionate depth
| Work | Delivery artifacts |
|---|---|
| Tiny | A single change.md with behavior, reuse, boundaries, acceptance, and risks |
| Standard | A brief design/plan.md and machine-checkable story specs |
| High risk | Detailed design, independent critique, QA, and human approval |
Knowledge-base and craftsmanship material are pulled in when a concrete problem needs them. SOUL is a reasoning guard across the four stages, not an extra stage.
Evidence before completion
Checks and reviews must apply to the current patch and acceptance inputs.
pr-ready establishes the required checks and review for the selected profile;
deployment and live host qualification need separate evidence.
Read the completion limits before making release claims.
The reference links to the canonical repository documentation, kernel commands, skill map, and GitHub Wiki overview. The README is the install-oriented project entry point; repository documents remain authoritative.