All skills
codenamev avatar

/architecture-review

@de7ccc8

Conducts a comprehensive multi-perspective architecture review using ALL architecture team members. Use when the user requests "Start architecture review", "Full architecture review", "Review architecture for version X.Y.Z", "Conduct comprehensive review", or when they want assessment from multiple perspectives. Do NOT use for single-specialist reviews (use specialist-review instead) or for status checks (use architecture-status instead).

Use this Skill: https://skilld.dev/gh/codenamev/ai-software-architect/architecture-review

This session only. Nothing lands on disk.

referencesreview-process.md

≈1.8k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Architecture Review Process - Detailed Guide

This document provides detailed instructions for conducting architecture reviews with all team members.

Table of Contents

  1. Individual Member Review Format
  2. Collaborative Discussion Process
  3. Analysis Guidelines by Review Type

Individual Member Review Format

Each member from .architecture/members.yml should review from their unique perspective using this structure:

Review Template

### [Member Name] - [Title]

**Perspective**: [Their unique viewpoint from members.yml]

#### Key Observations
- [Observation 1 - Specific finding about the system]
- [Observation 2 - Another specific finding]
- [Continue with notable observations]

#### Strengths
1. **[Strength Title]**: [Detailed description of what's working well and why it matters]
2. **[Strength Title]**: [Description]
3. [Continue with strengths - aim for 3-5 key strengths]

#### Concerns
1. **[Concern Title]** (Impact: High | Medium | Low)
   - **Issue**: [Clear description of what's wrong or could be improved]
   - **Why it matters**: [Impact on system, users, or development]
   - **Recommendation**: [Specific actionable recommendation]

2. **[Concern Title]** (Impact: High | Medium | Low)
   - **Issue**: [Description]
   - **Why it matters**: [Impact]
   - **Recommendation**: [Action]

[Continue with concerns - prioritize by impact]

#### Recommendations
1. **[Recommendation Title]** (Priority: High/Medium/Low, Effort: Small/Medium/Large)
   - **What**: [What to do]
   - **Why**: [Benefit or risk addressed]
   - **How**: [Brief implementation approach]

2. **[Recommendation Title]** (Priority: High/Medium/Low, Effort: Small/Medium/Large)
   - [Details]

[Continue with recommendations - 3-7 recommendations per member]

Member-Specific Guidance

Systems Architect:

  • Focus on overall system coherence and architectural patterns
  • Evaluate component interactions and integration points
  • Assess alignment with architectural principles
  • Consider long-term evolvability

Domain Expert:

  • Review how well architecture represents business domain
  • Evaluate bounded contexts and domain model accuracy
  • Check ubiquitous language usage
  • Assess semantic correctness

Security Specialist:

  • Identify security vulnerabilities and threats
  • Evaluate authentication, authorization, encryption
  • Review data protection and privacy considerations
  • Assess security boundaries and attack surface

Performance Specialist:

  • Identify performance bottlenecks
  • Evaluate scalability patterns
  • Review resource utilization
  • Assess caching, optimization opportunities

Maintainability Expert:

  • Evaluate code organization and clarity
  • Assess technical debt
  • Review complexity and coupling
  • Consider developer experience

AI Engineer (if applicable):

  • Review AI/ML integration patterns
  • Evaluate LLM application design
  • Assess agent orchestration
  • Review observability and evaluation

Pragmatic Enforcer (if pragmatic mode enabled):


Collaborative Discussion Process

After individual reviews, simulate a discussion between members to synthesize findings.

Discussion Structure

## Collaborative Discussion

**[Systems Architect]**: "[Opening statement about overall architecture]"

**[Domain Expert]**: "[Response or complementary view from domain perspective]"

**[Security Specialist]**: "[Security concerns or validation]"

[Continue discussion flow naturally]

### Common Ground

Team members agree on:
1. [Consensus point 1]
2. [Consensus point 2]
3. [Consensus point 3]

### Areas of Debate

**Topic: [Topic Title]**
- **[Member 1]**: [Their position]
- **[Member 2]**: [Their position]
- **Resolution**: [How team resolves or agrees to disagree]

### Priorities Established

The team agrees on these priorities:

**Critical (Address Immediately)**:
1. [Critical priority]
2. [Critical priority]

**Important (Address Soon)**:
1. [Important priority]
2. [Important priority]

**Nice-to-Have (Consider Later)**:
1. [Nice-to-have]
2. [Nice-to-have]

Discussion Best Practices

  1. Cross-Reference Findings: Members should reference and build on each other's observations
  2. Resolve Conflicts: When members disagree, discuss trade-offs and reach consensus
  3. Prioritize Together: Collaborate to rank recommendations by urgency and impact
  4. Be Realistic: Consider project constraints, deadlines, and team capacity
  5. Stay Constructive: Frame concerns as improvement opportunities

Analysis Guidelines by Review Type

Version Reviews

Focus on:

  • Overall architecture health at this milestone
  • Components and their interactions
  • Patterns and consistency
  • Technical debt accumulated
  • ADRs implemented or needed
  • Alignment with original architecture vision

Key Questions:

  • Is the architecture still coherent as system has evolved?
  • What technical debt needs addressing before next version?
  • Are architectural principles being followed?
  • What risks should be mitigated?

Feature Reviews

Focus on:

  • Feature implementation approach
  • Integration with existing architecture
  • Data flow and state management
  • Security implications
  • Performance impact
  • Test coverage

Key Questions:

  • Does this feature fit the existing architecture?
  • Are there better architectural approaches?
  • What are the integration risks?
  • How does this impact scalability?

Component Reviews

Focus on:

  • Component architecture and structure
  • Dependencies and coupling
  • Boundaries and interfaces
  • Responsibilities and cohesion
  • Testability

Key Questions:

  • Is component well-designed and focused?
  • Are boundaries clear and appropriate?
  • Is it properly decoupled?
  • Does it follow single responsibility principle?

Tips for High-Quality Reviews

Be Specific

  • ❌ "The code is messy"
  • ✅ "The UserService class has 15 methods mixing authentication, authorization, and profile management - violates single responsibility"

Provide Context

  • ❌ "Add caching"
  • ✅ "Add Redis caching for user profile queries - currently hitting DB 50+ times per page load causing 200ms delays"

Suggest Solutions

  • ❌ "Performance is bad"
  • ✅ "Batch database queries in OrderProcessor.process() to reduce N+1 queries - should improve processing time by 70%"

Balance Positive and Negative

  • Don't just list problems
  • Recognize what's working well
  • Explain why good patterns should be sustained

Be Actionable

  • Every concern should have a recommendation
  • Recommendations should be concrete and implementable
  • Estimate effort (Small/Medium/Large) and priority (High/Medium/Low)

Reference Documentation

For the full review document template, see ../assets/review-template.md.

For pragmatic mode integration, see pragmatic-integration.md.

Source: SKILL.md on GitHub

1 warning15d4 checks · Risk SAFE
  • Gen Agent Trust Hub15d

    The architecture-review skill is a well-structured orchestrator designed to manage multi-perspective architectural assessments. It follows security best practices by restricting tool access (specifically limiting Bash to git-only operations) and implementing explicit input sanitization. The skill delegates specialized tasks to subagents and aggregates findings using controlled templates, minimizing risks associated with processing external project data. No malicious patterns or security vulnerabilities were identified.

  • Socket15d

    No alerts

  • Snyk15d

    Risk: LOW · No issues

  • Runlayer7mo

    4/4 files flagged

Signed by skilld at de7ccc8. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 weeks ago.

Activeupdated 5 months ago
What it can do
Reads files Edits files Runs commands
All 6 allowed tools
ReadWriteGlobGrepBash(git:*)Agent

README badge

README badge for codenamev/ai-software-architect/architecture-review