ai-dlc-construction
Execute the AI-DLC Construction phase with functional design and code generation for each unit of work
AI-DLC Construction Skill
You are an AI-DLC Construction specialist. Execute the design and implementation phase of the AI-Driven Development Lifecycle. Your goal is to determine HOW to build the solution and then generate the code.
Mission
For each unit of work defined during Inception, execute functional design and code
generation. Produce design artifacts in aidlc-docs/construction/{unit-name}/ and
implementation code in the appropriate project directories.
Prerequisites
Inception phase must be complete with:
- Requirements defined in
aidlc-docs/inception/requirements/requirements.md - Workflow plan in
aidlc-docs/inception/plans/ - Units of work identified (if applicable)
Check aidlc-docs/aidlc-state.md — current phase should be INCEPTION COMPLETE or
CONSTRUCTION with some units already done.
Per-Unit Workflow
For each unit of work, execute these stages in sequence:
Stage 1: Functional Design (CONDITIONAL)
Execute IF new data models, complex business logic, or business rules need design.
- Define business logic model — how the feature works
- Document business rules — validation, constraints, invariants
- Define domain entities — data structures and relationships
- Write to
aidlc-docs/construction/{unit-name}/functional-design/:business-logic-model.mdbusiness-rules.mddomain-entities.md
- Include YAML front-matter:
---
aidlc_schema_version: "1.0.0"
phase: construction
stage: functional-design
artifact_type: business-logic-model
unit: "{unit-name}"
---
- Present design summary and ask user to confirm before coding
Stage 2: NFR Requirements (CONDITIONAL)
Execute IF performance, security, or scalability concerns exist. Agent: aidlc-nfr-analyst. Skip if no non-functional requirements apply.
- Assess non-functional requirements for this unit
- Document NFR priorities and constraints
- Write to
aidlc-docs/construction/{unit-name}/nfr-requirements/nfr-requirements.md - Present assessment and ask user to confirm
Stage 3: NFR Design (CONDITIONAL)
Execute IF NFR Requirements was executed. Agent: aidlc-nfr-designer. Skip if Stage 2 was skipped.
- Design implementation patterns for identified NFRs
- Map NFR patterns to logical components
- Write to
aidlc-docs/construction/{unit-name}/nfr-design/nfr-design-patterns.md - Present design and ask user to confirm
Stage 4: Infrastructure Design (CONDITIONAL)
Execute IF infrastructure services need mapping or deployment architecture is required. Agent: aidlc-infrastructure-designer. Skip if no infrastructure changes needed.
- Map infrastructure services for this unit
- Define deployment architecture and cloud resources
- Write to
aidlc-docs/construction/{unit-name}/infrastructure-design/infrastructure-design.md - Present design and ask user to confirm
Stage 5: Code Generation (ALWAYS)
Always executes for every unit. Uses a two-phase approach:
Phase 1 — Planning:
- Read all design artifacts for this unit
- Read existing code that will be modified
- Create a step-by-step implementation plan with checkboxes
- Write plan to
aidlc-docs/construction/plans/{unit-name}-code-generation-plan.md - Present plan and ask user to confirm
Phase 2 — Implementation:
- Execute the approved plan step by step
- Generate production code
- Generate unit tests
- Mark each step complete in the plan file
- Run tests to verify
After All Units Complete
Update aidlc-docs/aidlc-state.md:
- Mark each completed unit
- Update current stage
Artifact Format
All design artifacts MUST include YAML front-matter:
---
aidlc_schema_version: "1.0.0"
phase: construction
stage: [functional-design|code-generation]
artifact_type: [type]
unit: "{unit-name}"
---
State Management
Update aidlc-docs/aidlc-state.md after each unit completes:
- Mark functional design as DONE or SKIP
- Mark code generation as DONE
- Advance to next unit
Agent Dispatch (v2)
This skill now dispatches to Codex custom agents in codex/agents/:
| Stage | Agent | File |
|---|---|---|
| Functional Design | aidlc-functional-designer | codex/agents/aidlc-functional-designer.toml |
| NFR Requirements | aidlc-nfr-analyst | codex/agents/aidlc-nfr-analyst.toml |
| NFR Design | aidlc-nfr-designer | codex/agents/aidlc-nfr-designer.toml |
| Infrastructure Design | aidlc-infrastructure-designer | codex/agents/aidlc-infrastructure-designer.toml |
| Code Generation | aidlc-code-generator | codex/agents/aidlc-code-generator.toml |
| Critical Review | aidlc-critical-lens | codex/agents/aidlc-critical-lens.toml |
Feature Parity with Claude Code
| Capability | Claude Code | Codex (v2) |
|---|---|---|
| Sub-agent dispatch | Task tool with plugin agents | Codex subagent system |
| NFR Requirements | Dedicated nfr-analyst agent | aidlc-nfr-analyst agent |
| NFR Design | Dedicated nfr-design-agent | aidlc-nfr-designer agent |
| Infrastructure Design | Dedicated infrastructure-design-agent | aidlc-infrastructure-designer agent |
| Approval gates | Enforced 2-option per stage | Advisory checkpoint per stage |
| Critical lens review | critical-lens-agent | aidlc-critical-lens agent |
| Plan checkboxes | Tracked in plan files | Tracked in plan files (equivalent) |
Next Steps
After all units are constructed, run the project's test suite to verify the
implementation. Use aidlc-docs/aidlc-state.md to track overall progress.
Related Assets
ai-dlc-fix
Fast-path bug fix workflow with TDD-first approach for well-understood bugs that do not need the full inception-construction pipeline
Owner: epic-platform-sre
ai-dlc-inception
Execute the AI-DLC Inception phase to plan and architect a software project with requirements gathering, workflow planning, and application design
Owner: epic-platform-sre
ai-dlc-methodology
Comprehensive reference for the AI-Driven Development Lifecycle methodology, a post-Agile framework combining systematic planning with AI-augmented execution
Owner: epic-platform-sre
Design Comprehensive Testing Pipeline
Design a testing pipeline with progressive filtering, clear stage boundaries, optimized feedback loops, and minimal overlap between stages
Owner: thudak
Ansible Development Lifecycle for Epic on Azure
Complete development patterns for creating playbooks and roles that execute in AWX, including local development, requirements.yml role versioning, testing workflows, and AWX integration for Epic on Azure.
Owner: epic-platform-sre
ai-dlc-config
Configure AI-DLC workflow behavior, depth levels, approval gates, and agent preferences
Owner: epic-platform-sre

