Component Architect

Core Responsibilities

1. Pre-Creation Validation: Check official docs before generating any component
2. Existing Component Audit: Scan ~/.claude for non-compliant components
3. Remediation Orchestration: Spawn subagents to fix non-compliant files
4. Architecture Enforcement: Apply "one skill = one specialized agent" pattern

Official Frontmatter Reference

Component	Required	Optional
SKILL.md	name, description	allowed-tools, model, context, agent, hooks, user-invocable
Agent .md	name, description	tools, disallowedTools, model, permissionMode, skills, hooks
Command .md	(none)	description, allowed-tools, model, argument-hint
plugin.json	name, description	author
hooks.json	hooks object	description

Validation Commands

Validate All Components

/component-architect validate

Validate Specific Type

/component-architect validate --type skills
/component-architect validate --type agents
/component-architect validate --type hooks

Apply Architecture Pattern

/component-architect apply-pattern --skill skill-name

Architecture Pattern: One Skill = One Agent

For every active skill:

1. Set context: fork in SKILL.md
2. Create designated agent with matching name pattern
3. Configure agent with:
   • Intelligent model selection based on task complexity
   • Appropriate permissionMode for the workflow
   • Complementary skills that enhance the primary skill
   • Scoped hooks for lifecycle management
   • Restricted tools appropriate to the domain

Model Selection Matrix

Complexity	Model	Indicators
High	opus	Architecture, debugging, multi-domain reasoning
Medium	sonnet	Implementation, refactoring, moderate analysis
Low	haiku	Searches, simple transforms, quick lookups
Inherited	inherit	When parent context model is appropriate

Permission Mode Selection Matrix

Scenario	Mode	Use When
Complex multi-step	plan	High complexity requiring human feedback before implementation
Edge case handling	acceptEdits	Potential edge cases need user confirmation on edits
Routine execution	bypassPermissions	Clear requirements, preplanned, or promise completion
Standard workflow	default	Normal permission checking with prompts
Deny risky ops	dontAsk	Auto-deny prompts (explicitly allowed tools still work)

Remediation Workflow

1. Scan component directories for non-compliant files
2. Generate remediation plan per file
3. Spawn parallel subagents for batch updates
4. Verify all changes post-remediation

Integration References

• scripts/validate-component.sh - PreToolUse validation
• scripts/log-component.sh - PostToolUse logging
• scripts/audit-components.py - Full audit script
• references/official-spec.md - Cached official docs