Version: 1.0 Purpose: Comprehensive expert-driven quality checklists for common software delivery artifacts
This directory contains detailed checklists organized by expertise domain for common software artifacts. Each checklist is designed to be used by domain experts during artifact review.
| Domain | Abbreviation | Focus |
|---|---|---|
| Architecture | ARCH |
System design, patterns, modularity, scalability |
| Performance | PERF |
Efficiency, latency, throughput, resource optimization |
| Security | SEC |
Authentication, authorization, data protection, vulnerabilities |
| Reliability | REL |
Fault tolerance, error handling, recovery, resilience |
| Usability | UX |
User experience, accessibility, clarity, discoverability |
| Maintainability | MAINT |
Code quality, documentation, testability, technical debt |
| Compliance | COMPL |
Regulations, standards, audits, certifications |
| Data | DATA |
Data modeling, integrity, privacy, governance |
| Integration | INT |
APIs, third-party systems, interoperability |
| Operations | OPS |
Deployment, monitoring, observability, DevOps |
| Testing | TEST |
Test coverage, quality assurance, validation strategies |
| Business | BIZ |
Requirements alignment, value delivery, stakeholder needs |
| Artifact | Checklist File |
|---|---|
| PRD | PRD.md |
| DESIGN | DESIGN.md |
| ADR | ADR.md |
| FEATURE | FEATURE.md |
| CODING | CODING.md |
Each checklist contains:
Requirements that must be present in the artifact, organized by expertise domain.
Content that must not appear in the artifact, with guidance on where that content should be placed instead.
Use only these levels:
| Level | When to use | Expected action |
|---|---|---|
| CRITICAL | Missing/incorrect content makes the artifact unsafe, misleading, non-compliant, or not usable for downstream work. Typically blocks implementation or causes rework. | Must fix before continuing. |
| HIGH | Major quality gap that significantly increases risk/cost or makes important decisions ambiguous, but work can still proceed with caution. | Fix before approval/release. |
| MEDIUM | Meaningful improvement that increases clarity, testability, or maintainability; not usually blocking. | Fix when feasible; track if deferred. |
| LOW | Minor improvement, style/format polish, or optional enhancement. | Optional; address opportunistically. |
Rules:
- CRITICAL: Impacts correctness, safety, compliance, or makes requirements/design unverifiable.
- HIGH: Creates ambiguity in key flows, responsibilities, interfaces, or acceptance criteria.
- MEDIUM: Improves completeness/consistency; reduces future rework.
- LOW: Cosmetic or optional.
Example prompts for using checklists with an AI agent:
Review @docs/my-gear/PRD.md against @docs/checklists/PRD.md checklist.
Output findings in table format: Domain | Item | Severity | Finding | Recommendation
Review @docs/my-gear/DESIGN.md for Security (SEC) items only.
Use @docs/checklists/DESIGN.md as reference.
Scan @docs/my-gear/FEATURE.md against @docs/checklists/FEATURE.md.
Report only CRITICAL and HIGH severity issues.
Review @docs/my-gear/ADR.md against @docs/checklists/ADR.md.
For each finding, propose a concrete fix. Apply fixes directly if severity is MEDIUM or lower.
Review all artifacts in @docs/my-gear/ against corresponding checklists in @docs/checklists/.
Summarize by artifact, then by domain.
I'm about to commit changes to @docs/my-gear/PRD.md.
Quick check against @docs/checklists/PRD.md — any CRITICAL issues?
Based on @docs/checklists/DESIGN.md, identify missing sections in @docs/my-gear/DESIGN.md.
Generate draft content for missing CRITICAL items.
These checklists are integrated with the Constructor Studio PR review workflow. When reviewing PRs:
- PRD PRs: Use
PRD.md— covers requirements completeness, testability, traceability, and industry alignment - Design PRs: Use
DESIGN.md— covers architecture, trade-offs, API contracts, security, and antipatterns - ADR PRs: Use
ADR.md— covers decision significance, alternatives analysis, and overlap detection - Code PRs: Use
CODING.md— covers Rust correctness, architecture (ToolKit/SDK pattern), security (secure ORM), clippy/dylint compliance, testing, performance, etc.
The checklist is auto-selected by the /cf-gears-pr-review workflow based on the PR content. Configuration is in .cf-studio/config/pr-review.toml under the [[prompts]] entries.
Example prompts:
review PR 42
review PR #59 with CODE checklist
- Existing project docs (architecture docs, API specs, runbooks)
- Source code and tests
- Constructor Studio PR review config:
.cf-studio/config/pr-review.toml - PR review workflow:
studio-kit-gears/workflows/pr-review.md - PR status workflow:
studio-kit-gears/workflows/pr-status.md - PR review prompts:
studio-kit-gears/scripts/prompts/pr/ - PR review script:
studio-kit-gears/scripts/pr.py - PR review templates:
studio-kit-gears/artifacts/PR-CODE-REVIEW-TEMPLATE/ - PR review documentation:
docs/pr-review/