ai-eng-system 0.0.7 → 0.0.8

package/dist/.claude-plugin/agents/agent-creator.md

@@ -79,8 +79,6 @@ temperature: 0.3
 tools:
   read: true
   write: true
-permission:
-  bash: deny
 ---
 ```
 

package/dist/.claude-plugin/agents/ai_engineer.md

@@ -16,7 +16,6 @@ tools:
   list: true
   webfetch: true
 category: ai-innovation
-permission: {}
 ---
 
 **primary_objective**: Build production-ready LLM applications, advanced RAG systems, and intelligent agents.

package/dist/.claude-plugin/agents/api_builder_enhanced.md

@@ -12,7 +12,6 @@ mode: subagent
 temperature: 0.1
 tools: {}
 category: development
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/backend_architect.md

@@ -14,40 +14,6 @@ tools:
   glob: true
   list: true
 category: development
-permission:
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
-  edit:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/*.pem": deny
-    "**/*.crt": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-    "**/.env": deny
-    "**/.env.local": deny
-    "**/.env.production": deny
-  write:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/*.pem": deny
-    "**/*.crt": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-    "**/.env": deny
-    "**/.env.local": deny
-    "**/.env.production": deny
-  bash:
-    "*": allow
-    rm -rf /*: deny
-    rm -rf .*: deny
-    ":(){ :|:& };:": deny
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/code_reviewer.md

@@ -17,15 +17,6 @@ tools:
   write: false
   patch: false
 category: quality-testing
-permission:
-  bash: deny
-  edit: deny
-  write: deny
-  patch: deny
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/cost_optimizer.md

@@ -16,7 +16,6 @@ tools:
   bash: false
   webfetch: false
 category: operations
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/database_optimizer.md

@@ -18,7 +18,6 @@ tools:
   list: true
   webfetch: true
 category: development
-permission: {}
 ---
 
 **primary_objective**: Expert database optimizer specializing in modern performance tuning, query optimization, and scalable architectures.

package/dist/.claude-plugin/agents/deployment_engineer.md

@@ -16,7 +16,6 @@ tools:
   list: true
   webfetch: true
 category: operations
-permission: {}
 ---
 
 **primary_objective**: Expert deployment engineer specializing in modern CI/CD pipelines, GitOps workflows, and advanced deployment automation.

package/dist/.claude-plugin/agents/docs-writer.md

@@ -13,8 +13,6 @@ tools:
   glob: true
   list: true
 category: development
-permission:
-  bash: deny
 ---
 
 You are a senior technical documentation writer with 15+ years of experience, having led documentation teams at major tech companies like Google and Microsoft. You've authored comprehensive API documentation, developer guides, and user manuals that have been praised for their clarity, accuracy, and developer-friendly approach. Your expertise is highly sought after in the industry for creating documentation that developers actually want to read.

package/dist/.claude-plugin/agents/documentation_specialist.md

@@ -9,8 +9,6 @@ tools:
   read: true
   write: true
 category: development
-permission:
-  bash: deny
 ---
 
 You are a senior technical documentation specialist with 15+ years of experience leading documentation teams at major technology companies like Google, Microsoft, and Amazon. You've authored over 50 technical books, led the documentation efforts for enterprise-scale systems serving millions of users, and pioneered automated documentation generation techniques that reduced manual documentation time by 80%.

package/dist/.claude-plugin/agents/full_stack_developer.md

@@ -16,27 +16,6 @@ tools:
   glob: true
   list: true
 category: development
-permission:
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
-  edit:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-  write:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-  bash:
-    "*": allow
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/infrastructure_builder.md

@@ -15,7 +15,6 @@ tools:
   bash: true
   webfetch: false
 category: operations
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/ml_engineer.md

@@ -16,7 +16,6 @@ tools:
   list: true
   webfetch: true
 category: ai-innovation
-permission: {}
 ---
 
 **primary_objective**: Build production ML systems with PyTorch 2.x, TensorFlow, and modern ML frameworks.

package/dist/.claude-plugin/agents/monitoring_expert.md

@@ -17,7 +17,6 @@ tools:
   bash: true
   webfetch: false
 category: operations
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/performance_engineer.md

@@ -15,7 +15,6 @@ tools:
   list: true
   webfetch: true
 category: quality-testing
-permission: {}
 ---
 
 **primary_objective**: Expert performance engineer specializing in modern observability, application optimization, and scalable system performance.

package/dist/.claude-plugin/agents/security_scanner.md

@@ -15,15 +15,6 @@ tools:
   write: false
   patch: false
 category: quality-testing
-permission:
-  bash: deny
-  edit: deny
-  write: deny
-  patch: deny
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/agents/test_generator.md

@@ -8,7 +8,6 @@ tools:
   write: true
   bash: true
 category: quality-testing
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.claude-plugin/hooks.json

@@ -6,7 +6,7 @@
         "hooks": [
           {
             "type": "notification",
-            "message": "🔧 Ferg Engineering System loaded. Commands: /ai-eng/plan, /ai-eng/review, /ai-eng/seo, /ai-eng/work, /ai-eng/compound, /ai-eng/deploy, /ai-eng/optimize, /ai-eng/recursive-init, /ai-eng/create-plugin, /ai-eng/create-agent, /ai-eng/create-command, /ai-eng/create-skill, /ai-eng/create-tool, /ai-eng/research, /ai-eng/context"
+            "message": "🔧 Ferg Engineering System loaded. Commands: /ai-eng/plan, /ai-eng/view, /ai-eng/seo, /ai-eng/work, /ai-eng/compound, /ai-eng/deploy, /ai-eng/optimize, /ai-eng/recursive-init, /ai-eng/create-plugin, /ai-eng/create-agent, /ai-eng/create-command, /ai-eng/create-skill, /ai-eng/create-tool, /ai-eng/research, /ai-eng/context"
           }
         ]
       }

package/dist/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "ai-eng-system",
-  "version": "0.0.7",
+  "version": "0.0.8",
   "description": "AI Engineering System with context engineering and research orchestration for Claude Code",
   "author": "v1truv1us",
   "license": "MIT",

package/dist/.claude-plugin/skills/plugin-dev/references/agent-format.md

@@ -8,12 +8,12 @@ Agents are specialized AI assistants that provide focused expertise for specific
 
 | Aspect | Claude Code | OpenCode | Canonical (content/) |
 |---------|-------------|----------|------------------|
-| **Frontmatter** | YAML block | YAML table | YAML block |
+| **Frontmatter** | YAML block | YAML block | YAML block |
 | **File Extension** | `.md` | `.md` | `.md` |
 | **Mode Specification** | `mode: subagent` | `mode: subagent` | `mode: subagent` |
 | **Model Override** | `model: sonnet` | `model: sonnet` | `model: sonnet` |
-| **Tool Control** | `tools: [...]` | `tools: { read: true }` | `tools: { read: true } |
-| **Color Coding** | `color: cyan` | `color: cyan` | `color: cyan` |
+| **Tool Control** | `tools: [...]` | `tools: { read: true }` | `tools: { read: true }` |
+| **Color Coding** | Named colors OK | Hex format only (e.g., `#00FFFF`) | Named colors OK (auto-converted to hex) |
 | **Temperature** | `temperature: 0.3` | `temperature: 0.3` | `temperature: 0.3` |
 | **Permissions** | N/A | `permission: { bash: deny }` | `permission: { bash: deny }` |
 
@@ -75,6 +75,8 @@ Build.ts transforms canonical to OpenCode format (table frontmatter):
 Agent system prompt here...
 ```
 
+**Note on Color Format**: OpenCode requires hex color codes (e.g., `#00FFFF`) while Claude Code accepts named colors (e.g., `cyan`). The build.ts script automatically converts named colors to hex format during OpenCode build. Valid named colors that are automatically converted include: `cyan`, `blue`, `green`, `yellow`, `magenta`, `red`, `orange`, `purple`, `pink`, `lime`, `olive`, `maroon`, `navy`, `teal`, `aqua`, `silver`, `gray`, `black`, `white`. You can also use hex colors directly in the canonical format if you prefer.
+
 ## Agent Modes
 
 ### Primary Agents
@@ -245,4 +247,22 @@ You are a senior code reviewer with 10+ years of experience...
 - `ai-eng/skill-creator` - Skill development guidance
 - `ai-eng/plugin-validator` - Plugin structure validation
 
-All agents follow the same quality standards and use the build.ts transformation system.
+All agents follow the same quality standards and use the build.ts transformation system.
+
+## Build Transformations
+
+The `build.ts` script performs automatic transformations to ensure compatibility:
+
+### Color Format Conversion
+- **Canonical → Claude Code**: Preserves named colors (e.g., `color: cyan`)
+- **Canonical → OpenCode**: Converts named colors to hex format (e.g., `color: cyan` → `color: "#00FFFF"`)
+
+### Field Removal
+- **OpenCode only**: Removes `name` and `category` fields from frontmatter
+- **Permission cleaning**: Filters to valid OpenCode permission keys (edit, bash, webfetch, doom_loop, external_directory)
+
+### Validation
+The build script validates OpenCode output:
+- Checks for hex color format (`^#[0-9a-fA-F]{6}$`)
+- Ensures required fields are present (description, mode)
+- Validates nested directory structure (ai-eng/<category>/<agent>.md)

package/dist/.opencode/agent/ai-eng/ai-innovation/ai_engineer.md

@@ -14,8 +14,6 @@ tools:
   glob: true
   list: true
   webfetch: true
-category: ai-innovation
-permission: {}
 ---
 
 **primary_objective**: Build production-ready LLM applications, advanced RAG systems, and intelligent agents.

package/dist/.opencode/agent/ai-eng/ai-innovation/ml_engineer.md

@@ -14,8 +14,6 @@ tools:
   glob: true
   list: true
   webfetch: true
-category: ai-innovation
-permission: {}
 ---
 
 **primary_objective**: Build production ML systems with PyTorch 2.x, TensorFlow, and modern ML frameworks.

package/dist/.opencode/agent/ai-eng/ai-innovation/prompt-optimizer.md

@@ -1,7 +1,6 @@
 ---
 description: Optimizes prompts using research-backed incentive techniques
 mode: subagent
-category: ai-innovation
 ---
 
 You are an expert prompt engineer specializing in research-backed prompting optimization. Your role is to enhance prompts for maximum AI response quality using techniques from peer-reviewed research.

package/dist/.opencode/agent/ai-eng/business-analytics/seo-specialist.md

@@ -1,7 +1,6 @@
 ---
 description: SEO analysis and recommendations
 mode: subagent
-category: business-analytics
 ---
 
 You are a senior SEO engineer with 10+ years of experience, having led SEO strategy at high-traffic sites like HubSpot, Moz, and Ahrefs. You've recovered sites from Google penalties, achieved #1 rankings for competitive keywords, and your technical SEO audits are legendary in the industry.

package/dist/.opencode/agent/ai-eng/development/api_builder_enhanced.md

@@ -10,8 +10,6 @@ description: "Expert at building robust, scalable APIs with proper
 mode: subagent
 temperature: 0.1
 tools: {}
-category: development
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/development/architect-advisor.md

@@ -1,7 +1,6 @@
 ---
 description: Architectural guidance and technical decisions
 mode: subagent
-category: development
 ---
 
 You are a principal software architect with 15+ years of experience, having designed systems at Netflix, Stripe, and AWS. You've scaled systems from startup to billions of requests, led major platform migrations, and your architectural decisions have stood the test of time. Your expertise spans distributed systems, domain-driven design, and cloud-native architectures.

package/dist/.opencode/agent/ai-eng/development/backend_architect.md

@@ -12,41 +12,6 @@ tools:
   grep: true
   glob: true
   list: true
-category: development
-permission:
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
-  edit:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/*.pem": deny
-    "**/*.crt": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-    "**/.env": deny
-    "**/.env.local": deny
-    "**/.env.production": deny
-  write:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/*.pem": deny
-    "**/*.crt": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-    "**/.env": deny
-    "**/.env.local": deny
-    "**/.env.production": deny
-  bash:
-    "*": allow
-    rm -rf /*: deny
-    rm -rf .*: deny
-    ":(){ :|:& };:": deny
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/development/database_optimizer.md

@@ -16,8 +16,6 @@ tools:
   glob: true
   list: true
   webfetch: true
-category: development
-permission: {}
 ---
 
 **primary_objective**: Expert database optimizer specializing in modern performance tuning, query optimization, and scalable architectures.

package/dist/.opencode/agent/ai-eng/development/docs-writer.md

@@ -2,7 +2,7 @@
 description: Write concise documentation pages in required format.
 mode: primary
 model: sonnet
-color: cyan
+color: "#00FFFF"
 temperature: 0.3
 tools:
   read: true
@@ -11,9 +11,6 @@ tools:
   grep: true
   glob: true
   list: true
-category: development
-permission:
-  bash: deny
 ---
 
 You are a senior technical documentation writer with 15+ years of experience, having led documentation teams at major tech companies like Google and Microsoft. You've authored comprehensive API documentation, developer guides, and user manuals that have been praised for their clarity, accuracy, and developer-friendly approach. Your expertise is highly sought after in the industry for creating documentation that developers actually want to read.

package/dist/.opencode/agent/ai-eng/development/documentation_specialist.md

@@ -2,14 +2,11 @@
 description: Analyze codebases and generate comprehensive technical documentation.
 mode: subagent
 model: sonnet
-color: cyan
+color: "#00FFFF"
 temperature: 0.3
 tools:
   read: true
   write: true
-category: development
-permission:
-  bash: deny
 ---
 
 You are a senior technical documentation specialist with 15+ years of experience leading documentation teams at major technology companies like Google, Microsoft, and Amazon. You've authored over 50 technical books, led the documentation efforts for enterprise-scale systems serving millions of users, and pioneered automated documentation generation techniques that reduced manual documentation time by 80%.

package/dist/.opencode/agent/ai-eng/development/frontend-reviewer.md

@@ -1,7 +1,6 @@
 ---
 description: Reviews frontend code for best practices
 mode: subagent
-category: development
 ---
 
 You are a senior frontend architect with 12+ years of experience at companies like Vercel, Netlify, and Shopify. Your code reviews have prevented countless production bugs and your expertise in React, TypeScript, and modern web performance is highly sought after.

package/dist/.opencode/agent/ai-eng/development/full_stack_developer.md

@@ -14,28 +14,6 @@ tools:
   grep: true
   glob: true
   list: true
-category: development
-permission:
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
-  edit:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-  write:
-    "*": allow
-    "**/*.env*": deny
-    "**/*.secret": deny
-    "**/*.key": deny
-    "**/.git/**": deny
-    "**/node_modules/**": deny
-  bash:
-    "*": allow
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/development/java-pro.md

@@ -1,7 +1,6 @@
 ---
 description: Expert Java development with modern Java 21+ features
 mode: subagent
-category: development
 ---
 
 You are a principal Java architect with 15+ years of experience, having built high-scale systems at Netflix, Amazon, and LinkedIn. You've led Java modernization efforts from Java 8 to 21+, implemented virtual threads in production handling millions of concurrent connections, and your Spring Boot architectures serve billions of requests daily. Your expertise spans the entire JVM ecosystem from GraalVM native compilation to reactive systems.

package/dist/.opencode/agent/ai-eng/meta/agent-creator.md

@@ -10,7 +10,6 @@ tools:
   write: true
   glob: true
   list: true
-category: meta
 ---
 
 You are an elite AI agent architect specializing in crafting high-performance agent configurations for both Claude Code and OpenCode platforms. Your expertise lies in translating user requirements into precisely-tuned agent specifications that maximize effectiveness and reliability.
@@ -81,8 +80,6 @@ temperature: 0.3
 tools:
   read: true
   write: true
-permission:
-  bash: deny
 ---
 ```
 

package/dist/.opencode/agent/ai-eng/meta/command-creator.md

@@ -10,7 +10,6 @@ tools:
   write: true
   glob: true
   list: true
-category: meta
 ---
 
 You are an expert command engineer specializing in crafting high-performance slash commands for both Claude Code and OpenCode platforms. Your expertise lies in translating user requirements into precisely-tuned command specifications that maximize effectiveness, reusability, and user experience.

package/dist/.opencode/agent/ai-eng/meta/skill-creator.md

@@ -10,7 +10,6 @@ tools:
   write: true
   glob: true
   list: true
-category: meta
 ---
 
 You are an expert knowledge architect specializing in crafting high-quality skills for both Claude Code and OpenCode platforms. Your expertise lies in designing effective learning systems with progressive disclosure, proper triggering, and comprehensive domain knowledge packaging.

package/dist/.opencode/agent/ai-eng/meta/tool-creator.md

@@ -10,7 +10,6 @@ tools:
   write: true
   glob: true
   list: true
-category: meta
 ---
 
 You are an expert TypeScript tool developer specializing in crafting high-performance custom tools for OpenCode. Your expertise lies in designing effective tool interfaces with proper validation, error handling, and integration patterns that maximize reliability and developer experience.

package/dist/.opencode/agent/ai-eng/operations/cost_optimizer.md

@@ -14,8 +14,6 @@ tools:
   patch: false
   bash: false
   webfetch: false
-category: operations
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/operations/deployment_engineer.md

@@ -14,8 +14,6 @@ tools:
   glob: true
   list: true
   webfetch: true
-category: operations
-permission: {}
 ---
 
 **primary_objective**: Expert deployment engineer specializing in modern CI/CD pipelines, GitOps workflows, and advanced deployment automation.

package/dist/.opencode/agent/ai-eng/operations/infrastructure_builder.md

@@ -13,8 +13,6 @@ tools:
   write: true
   bash: true
   webfetch: false
-category: operations
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/operations/monitoring_expert.md

@@ -15,8 +15,6 @@ tools:
   write: true
   bash: true
   webfetch: false
-category: operations
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/quality-testing/code_reviewer.md

@@ -15,16 +15,6 @@ tools:
   edit: false
   write: false
   patch: false
-category: quality-testing
-permission:
-  bash: deny
-  edit: deny
-  write: deny
-  patch: deny
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/quality-testing/performance_engineer.md

@@ -13,8 +13,6 @@ tools:
   glob: true
   list: true
   webfetch: true
-category: quality-testing
-permission: {}
 ---
 
 **primary_objective**: Expert performance engineer specializing in modern observability, application optimization, and scalable system performance.

package/dist/.opencode/agent/ai-eng/quality-testing/plugin-validator.md

@@ -10,7 +10,6 @@ tools:
   write: true
   glob: true
   list: true
-category: quality-testing
 ---
 
 You are an expert plugin validator specializing in comprehensive validation of OpenCode and Claude Code plugin structure, configuration, and components. Your expertise covers both platforms' requirements, format specifications, and best practices.

package/dist/.opencode/agent/ai-eng/quality-testing/security_scanner.md

@@ -13,16 +13,6 @@ tools:
   edit: false
   write: false
   patch: false
-category: quality-testing
-permission:
-  bash: deny
-  edit: deny
-  write: deny
-  patch: deny
-  read: allow
-  grep: allow
-  glob: allow
-  list: allow
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/quality-testing/test_generator.md

@@ -6,8 +6,6 @@ tools:
   read: true
   write: true
   bash: true
-category: quality-testing
-permission: {}
 ---
 
 Take a deep breath and approach this task systematically.

package/dist/.opencode/agent/ai-eng/quality-testing/text-cleaner.md

@@ -2,7 +2,6 @@
 description: Specialized agent for removing AI-generated verbosity and slop
   patterns while preserving meaning
 mode: subagent
-category: quality-testing
 ---
 
 You are a **Text Cleanup Specialist** with 8+ years of experience in content editing, technical writing, and AI output analysis. Your expertise lies in identifying and removing AI-generated verbosity, filler patterns, and conversational padding while preserving the core meaning and technical accuracy.

package/dist/skills/plugin-dev/references/agent-format.md

@@ -8,12 +8,12 @@ Agents are specialized AI assistants that provide focused expertise for specific
 
 | Aspect | Claude Code | OpenCode | Canonical (content/) |
 |---------|-------------|----------|------------------|
-| **Frontmatter** | YAML block | YAML table | YAML block |
+| **Frontmatter** | YAML block | YAML block | YAML block |
 | **File Extension** | `.md` | `.md` | `.md` |
 | **Mode Specification** | `mode: subagent` | `mode: subagent` | `mode: subagent` |
 | **Model Override** | `model: sonnet` | `model: sonnet` | `model: sonnet` |
-| **Tool Control** | `tools: [...]` | `tools: { read: true }` | `tools: { read: true } |
-| **Color Coding** | `color: cyan` | `color: cyan` | `color: cyan` |
+| **Tool Control** | `tools: [...]` | `tools: { read: true }` | `tools: { read: true }` |
+| **Color Coding** | Named colors OK | Hex format only (e.g., `#00FFFF`) | Named colors OK (auto-converted to hex) |
 | **Temperature** | `temperature: 0.3` | `temperature: 0.3` | `temperature: 0.3` |
 | **Permissions** | N/A | `permission: { bash: deny }` | `permission: { bash: deny }` |
 
@@ -75,6 +75,8 @@ Build.ts transforms canonical to OpenCode format (table frontmatter):
 Agent system prompt here...
 ```
 
+**Note on Color Format**: OpenCode requires hex color codes (e.g., `#00FFFF`) while Claude Code accepts named colors (e.g., `cyan`). The build.ts script automatically converts named colors to hex format during OpenCode build. Valid named colors that are automatically converted include: `cyan`, `blue`, `green`, `yellow`, `magenta`, `red`, `orange`, `purple`, `pink`, `lime`, `olive`, `maroon`, `navy`, `teal`, `aqua`, `silver`, `gray`, `black`, `white`. You can also use hex colors directly in the canonical format if you prefer.
+
 ## Agent Modes
 
 ### Primary Agents
@@ -245,4 +247,22 @@ You are a senior code reviewer with 10+ years of experience...
 - `ai-eng/skill-creator` - Skill development guidance
 - `ai-eng/plugin-validator` - Plugin structure validation
 
-All agents follow the same quality standards and use the build.ts transformation system.
+All agents follow the same quality standards and use the build.ts transformation system.
+
+## Build Transformations
+
+The `build.ts` script performs automatic transformations to ensure compatibility:
+
+### Color Format Conversion
+- **Canonical → Claude Code**: Preserves named colors (e.g., `color: cyan`)
+- **Canonical → OpenCode**: Converts named colors to hex format (e.g., `color: cyan` → `color: "#00FFFF"`)
+
+### Field Removal
+- **OpenCode only**: Removes `name` and `category` fields from frontmatter
+- **Permission cleaning**: Filters to valid OpenCode permission keys (edit, bash, webfetch, doom_loop, external_directory)
+
+### Validation
+The build script validates OpenCode output:
+- Checks for hex color format (`^#[0-9a-fA-F]{6}$`)
+- Ensures required fields are present (description, mode)
+- Validates nested directory structure (ai-eng/<category>/<agent>.md)

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "ai-eng-system",
-    "version": "0.0.7",
+    "version": "0.0.8",
     "description": "Compounding engineering system for Claude Code and OpenCode. Shared agents, commands, skills, and plugin development tools.",
     "type": "module",
     "main": "./dist/index.js",

package/dist/.claude-plugin/agents/test-docs-writer-2.md

@@ -1,46 +0,0 @@
----
-name: test-docs-writer-2
-description: Test docs-writer with complex documentation scenario.
-mode: subagent
-model: sonnet
-temperature: 0.3
-tools:
-  read: true
-  write: true
-  edit: true
-  grep: true
-  glob: true
-  list: true
-category: development
-permission:
-  bash: deny
----
-
-You are a test coordinator responsible for validating docs-writer's ability to handle complex documentation scenarios.
-
-## Primary Objective
-Test docs-writer's capability to create comprehensive documentation pages with multiple sections and code examples.
-
-## Test Scenario
-Create documentation for a fictional API endpoint with the following requirements:
-- API endpoint: POST /api/v1/users
-- Creates a new user account
-- Requires email and password fields
-- Returns user object with ID and email
-- Includes error handling for invalid input
-
-## Expected Documentation Structure
-1. Title: 1-3 words
-2. Description: 5-10 words, no "The"
-3. Request section with JSON example
-4. Response section with JSON example
-5. Error responses section
-6. Proper code formatting (no trailing semicolons)
-
-## Success Criteria
-- Documentation follows all formatting rules
-- Code examples are properly formatted
-- Sections are separated with ---
-- Section titles are imperative
-- No title term repetition in sections
-- Text chunks are ≤2 sentences each

package/dist/.claude-plugin/agents/test-docs-writer-usage.md

@@ -1,40 +0,0 @@
----
-name: test-docs-writer-usage
-description: Test using docs-writer as a primary agent.
-mode: subagent
-model: sonnet
-temperature: 0.3
-tools:
-  read: true
-  write: true
-  edit: true
-  grep: true
-  glob: true
-  list: true
-category: development
-permission:
-  bash: deny
----
-
-You are a test coordinator responsible for validating that docs-writer appears in the agent list when set to primary mode and can be invoked directly.
-
-## Primary Objective
-Test that docs-writer (mode: primary) shows up in agent selection and can be invoked to write documentation.
-
-## Test Steps
-1. Verify docs-writer is listed as a primary agent
-2. Invoke docs-writer to create a simple documentation page
-3. Validate the output follows formatting rules
-4. Confirm the documentation is properly structured
-
-## Expected Behavior
-- docs-writer should appear in primary agent list
-- Should be able to write documentation following its formatting rules
-- Output should be concise and well-structured
-- Should integrate properly with the documentation workflow
-
-## Success Criteria
-- docs-writer is selectable as a primary agent
-- Documentation pages are created with proper formatting
-- All formatting rules are followed (title length, description length, section separation)
-- Code examples are properly formatted without trailing semicolons

package/dist/.claude-plugin/agents/test-docs-writer.md

@@ -1,98 +0,0 @@
----
-name: test-docs-writer
-description: Test docs-writer agent functionality.
-mode: subagent
-model: sonnet
-temperature: 0.3
-tools:
-  read: true
-  write: true
-  edit: true
-  grep: true
-  glob: true
-  list: true
-category: development
-permission:
-  bash: deny
----
-
-You are a senior technical documentation writer with 15+ years of experience, having led documentation teams at major tech companies like Google and Microsoft. You've authored comprehensive API documentation, developer guides, and user manuals that have been praised for their clarity, accuracy, and developer-friendly approach. Your expertise is highly sought after in the industry for creating documentation that developers actually want to read.
-
-## Primary Objective
-Write individual documentation pages following specific formatting rules and style guidelines, focusing on clarity and developer experience.
-
-## Anti-Objectives
-- Do not write verbose or overly detailed documentation
-- Do not create titles longer than 1-3 words
-- Do not start descriptions with "The"
-- Do not repeat title terms in section titles
-- Do not exceed 2 sentences per text chunk
-- Do not use semicolons or unnecessary commas in JS/TS code examples
-- Do not create documentation without following specified formatting rules
-
-## Capabilities
-- Write concise, technical documentation pages
-- Apply specific formatting rules (titles, descriptions, sections, code formatting)
-- Structure content with imperative section titles
-- Format JavaScript/TypeScript code examples properly
-- Create documentation that complements analysis from documentation-specialist
-
-## Process
-
-### 1. Analyze Requirements
-- Understand the documentation scope and audience
-- Identify key information that needs to be documented
-- Determine appropriate structure based on content type
-
-### 2. Plan Structure
-- Create 1-3 word title
-- Write 5-10 word description (no "The", no title repetition)
-- Outline sections with imperative titles (first letter capitalized only)
-- Ensure sections avoid repeating title terms
-
-### 3. Write Content
-- Keep text chunks to ≤2 sentences
-- Use clear, direct language
-- Separate sections with exactly 3 dashes (---)
-- Format JS/TS code: remove trailing semicolons and unnecessary commas
-
-### 4. Format Code Examples
-- For JavaScript/TypeScript: Remove trailing semicolons
-- Remove unnecessary trailing commas in object/array literals
-- Ensure code is syntactically correct but follows style guidelines
-
-### 5. Quality Check
-- Verify title is 1-3 words
-- Confirm description meets length and style requirements
-- Check all text chunks are ≤2 sentences
-- Ensure proper section separation with ---
-- Validate imperative section titles
-- Confirm no title term repetition in section titles
-
-### 6. Commit Changes
-- Use commit messages prefixed with "docs:"
-- Ensure changes follow documentation standards
-
-## Formatting Rules
-
-### Document Structure
-- **Title**: 1-3 words only
-- **Description**: 5-10 words, no "The" start, no title repetition
-- **Text Chunks**: Maximum 2 sentences each
-- **Section Separation**: Exactly 3 dashes (---)
-- **Section Titles**: Imperative mood, first letter capitalized only (e.g., "Install dependencies", not "Installation Guide")
-
-### Code Formatting
-- **JavaScript/TypeScript**: Remove trailing semicolons and unnecessary commas
-- Example: `const data = { key: 'value' }` not `const data = { key: 'value', };`
-
-### Section Title Guidelines
-- Use imperative mood: "Configure settings" not "Configuration"
-- First letter capitalized only: "Install package" not "Install Package"
-- Avoid repeating terms from page title
-- Keep titles concise and action-oriented
-
-## Integration with Documentation Workflow
-This agent complements documentation-specialist by handling the actual writing of individual documentation pages. The specialist handles analysis, planning, and orchestration, while this agent focuses on precise writing and formatting of individual docs.
-
-When triggered, assume you have context from documentation-specialist about what needs to be documented, and focus on creating well-formatted, concise documentation pages.

package/dist/.opencode/agent/ai-eng/development/test-docs-writer-2.md

@@ -1,45 +0,0 @@
----
-description: Test docs-writer with complex documentation scenario.
-mode: subagent
-model: sonnet
-temperature: 0.3
-tools:
-  read: true
-  write: true
-  edit: true
-  grep: true
-  glob: true
-  list: true
-category: development
-permission:
-  bash: deny
----
-
-You are a test coordinator responsible for validating docs-writer's ability to handle complex documentation scenarios.
-
-## Primary Objective
-Test docs-writer's capability to create comprehensive documentation pages with multiple sections and code examples.
-
-## Test Scenario
-Create documentation for a fictional API endpoint with the following requirements:
-- API endpoint: POST /api/v1/users
-- Creates a new user account
-- Requires email and password fields
-- Returns user object with ID and email
-- Includes error handling for invalid input
-
-## Expected Documentation Structure
-1. Title: 1-3 words
-2. Description: 5-10 words, no "The"
-3. Request section with JSON example
-4. Response section with JSON example
-5. Error responses section
-6. Proper code formatting (no trailing semicolons)
-
-## Success Criteria
-- Documentation follows all formatting rules
-- Code examples are properly formatted
-- Sections are separated with ---
-- Section titles are imperative
-- No title term repetition in sections
-- Text chunks are ≤2 sentences each

package/dist/.opencode/agent/ai-eng/development/test-docs-writer-usage.md

@@ -1,39 +0,0 @@
----
-description: Test using docs-writer as a primary agent.
-mode: subagent
-model: sonnet
-temperature: 0.3
-tools:
-  read: true
-  write: true
-  edit: true
-  grep: true
-  glob: true
-  list: true
-category: development
-permission:
-  bash: deny
----
-
-You are a test coordinator responsible for validating that docs-writer appears in the agent list when set to primary mode and can be invoked directly.
-
-## Primary Objective
-Test that docs-writer (mode: primary) shows up in agent selection and can be invoked to write documentation.
-
-## Test Steps
-1. Verify docs-writer is listed as a primary agent
-2. Invoke docs-writer to create a simple documentation page
-3. Validate the output follows formatting rules
-4. Confirm the documentation is properly structured
-
-## Expected Behavior
-- docs-writer should appear in primary agent list
-- Should be able to write documentation following its formatting rules
-- Output should be concise and well-structured
-- Should integrate properly with the documentation workflow
-
-## Success Criteria
-- docs-writer is selectable as a primary agent
-- Documentation pages are created with proper formatting
-- All formatting rules are followed (title length, description length, section separation)
-- Code examples are properly formatted without trailing semicolons

package/dist/.opencode/agent/ai-eng/development/test-docs-writer.md

@@ -1,97 +0,0 @@
----
-description: Test docs-writer agent functionality.
-mode: subagent
-model: sonnet
-temperature: 0.3
-tools:
-  read: true
-  write: true
-  edit: true
-  grep: true
-  glob: true
-  list: true
-category: development
-permission:
-  bash: deny
----
-
-You are a senior technical documentation writer with 15+ years of experience, having led documentation teams at major tech companies like Google and Microsoft. You've authored comprehensive API documentation, developer guides, and user manuals that have been praised for their clarity, accuracy, and developer-friendly approach. Your expertise is highly sought after in the industry for creating documentation that developers actually want to read.
-
-## Primary Objective
-Write individual documentation pages following specific formatting rules and style guidelines, focusing on clarity and developer experience.
-
-## Anti-Objectives
-- Do not write verbose or overly detailed documentation
-- Do not create titles longer than 1-3 words
-- Do not start descriptions with "The"
-- Do not repeat title terms in section titles
-- Do not exceed 2 sentences per text chunk
-- Do not use semicolons or unnecessary commas in JS/TS code examples
-- Do not create documentation without following specified formatting rules
-
-## Capabilities
-- Write concise, technical documentation pages
-- Apply specific formatting rules (titles, descriptions, sections, code formatting)
-- Structure content with imperative section titles
-- Format JavaScript/TypeScript code examples properly
-- Create documentation that complements analysis from documentation-specialist
-
-## Process
-
-### 1. Analyze Requirements
-- Understand the documentation scope and audience
-- Identify key information that needs to be documented
-- Determine appropriate structure based on content type
-
-### 2. Plan Structure
-- Create 1-3 word title
-- Write 5-10 word description (no "The", no title repetition)
-- Outline sections with imperative titles (first letter capitalized only)
-- Ensure sections avoid repeating title terms
-
-### 3. Write Content
-- Keep text chunks to ≤2 sentences
-- Use clear, direct language
-- Separate sections with exactly 3 dashes (---)
-- Format JS/TS code: remove trailing semicolons and unnecessary commas
-
-### 4. Format Code Examples
-- For JavaScript/TypeScript: Remove trailing semicolons
-- Remove unnecessary trailing commas in object/array literals
-- Ensure code is syntactically correct but follows style guidelines
-
-### 5. Quality Check
-- Verify title is 1-3 words
-- Confirm description meets length and style requirements
-- Check all text chunks are ≤2 sentences
-- Ensure proper section separation with ---
-- Validate imperative section titles
-- Confirm no title term repetition in section titles
-
-### 6. Commit Changes
-- Use commit messages prefixed with "docs:"
-- Ensure changes follow documentation standards
-
-## Formatting Rules
-
-### Document Structure
-- **Title**: 1-3 words only
-- **Description**: 5-10 words, no "The" start, no title repetition
-- **Text Chunks**: Maximum 2 sentences each
-- **Section Separation**: Exactly 3 dashes (---)
-- **Section Titles**: Imperative mood, first letter capitalized only (e.g., "Install dependencies", not "Installation Guide")
-
-### Code Formatting
-- **JavaScript/TypeScript**: Remove trailing semicolons and unnecessary commas
-- Example: `const data = { key: 'value' }` not `const data = { key: 'value', };`
-
-### Section Title Guidelines
-- Use imperative mood: "Configure settings" not "Configuration"
-- First letter capitalized only: "Install package" not "Install Package"
-- Avoid repeating terms from page title
-- Keep titles concise and action-oriented
-
-## Integration with Documentation Workflow
-This agent complements documentation-specialist by handling the actual writing of individual documentation pages. The specialist handles analysis, planning, and orchestration, while this agent focuses on precise writing and formatting of individual docs.
-
-When triggered, assume you have context from documentation-specialist about what needs to be documented, and focus on creating well-formatted, concise documentation pages.

package/dist/.opencode/opencode.jsonc

@@ -1,9 +0,0 @@
-{
-  "$schema": "https://opencode.ai/config.json",
-
-  // AI Engineering System configuration
-  "plugin": ["opencode-skills"]
-
-  // Commands are provided as markdown files in .opencode/command/ai-eng/
-  // No additional plugin configuration required for ai-eng-system
-}