ai-workflow-init 6.2.3 → 6.3.2

package/docs/sync-workflow/factory-droid.md

@@ -0,0 +1,234 @@
+# Factory Droid Sync Guide
+
+Sync Claude Code workflows to Factory Droid format.
+
+---
+
+## Prerequisites
+
+- Source: `.claude/commands/*.md`, `.claude/skills/**/SKILL.md`, `.claude/CLAUDE.md`
+- Target: `.factory/commands/*.md`, `.factory/skills/*/SKILL.md`, `.factory/droids/*.md`
+
+---
+
+## Step 1: Fetch Factory Droid Documentation
+
+**Search queries (try in order until successful):**
+1. `site:docs.factory.ai cli configuration commands skills`
+2. `factory.ai droid custom commands skills documentation`
+3. `factory droid cli configuration custom-droids`
+
+**Key documentation URLs:**
+- `https://docs.factory.ai/cli/configuration/custom-slash-commands` - Custom commands
+- `https://docs.factory.ai/cli/configuration/skills` - Skills configuration
+- `https://docs.factory.ai/cli/configuration/custom-droids` - Custom droids (sub-agents)
+
+---
+
+## Step 2: Conversion Rules
+
+### Commands (`.factory/commands/*.md`)
+
+**Location**: `.factory/commands/{name}.md` (workspace) or `~/.factory/commands/` (personal)
+
+**File format:** Markdown (`.md`) or executable with shebang (`#!`)
+
+**Frontmatter:**
+- `description`: command description
+- `argument-hint`: optional hint for arguments (e.g., `<feature-name>`)
+
+**Arguments:**
+- `$ARGUMENTS` for all text after command name
+- Executable commands receive `$1`, `$2`, etc.
+
+**Tool references conversion:**
+| Claude Code | Factory Droid |
+|-------------|---------------|
+| `AskUserQuestion(...)` | (handled by TUI) |
+| `Task(subagent_type='Explore')` | Search codebase |
+| `Task(subagent_type='General')` | General analysis |
+| `Read(file_path=...)` | Read file |
+| `Write(file_path=...)` | Write file |
+| `Edit(file_path=...)` | Edit file |
+| `Glob(pattern=...)` | Find files |
+| `Grep(pattern=...)` | Search content |
+| `Bash(command=...)` | Run command |
+
+### Skills (`.factory/skills/<name>/SKILL.md`)
+
+**Location**: `.factory/skills/{skill-name}/SKILL.md` (workspace) or `~/.factory/skills/` (personal)
+
+**File format:** `SKILL.md` or `skill.mdx`
+
+**Frontmatter (required):**
+- `name`: skill name (should match directory name)
+- `description`: clear description of when to use the skill
+
+**Content:** Markdown instructions for skill behavior
+
+### Droids (`.factory/droids/*.md`)
+
+**Location**: `.factory/droids/{droid-name}.md` (project) or `~/.factory/droids/` (personal)
+
+**File format:** Markdown with YAML frontmatter and non-empty body
+
+**Frontmatter:**
+- `name`: Required, droid identifier (matches filename)
+- `description`: Optional, shown in UI
+- `model`: `inherit` (use default) or specific model identifier
+- `reasoningEffort`: Optional (`low`, `medium`, `high`)
+- `tools`: Array of tool IDs or category string
+
+**Tool categories:**
+- `read-only` - Read, LS, Grep, Glob
+- `edit` - Write, Edit
+- `execute` - Bash, shell commands
+- `web` - WebSearch, WebFetch
+- `mcp` - MCP tools
+
+**Note:** `TodoWrite` tool is automatic for all droids.
+
+---
+
+## Step 3: Templates
+
+### Command Template
+
+```markdown
+---
+description: {description}
+argument-hint: {optional hint for arguments, e.g., "<feature-name>"}
+---
+
+## Goal
+
+{goal from Claude command}
+
+## Instructions
+
+{step-by-step instructions}
+
+### Step 1: {step name}
+
+{converted step content}
+- Search codebase for relevant files
+- Read and analyze existing code
+- Write/edit files as needed
+
+...
+
+## Notes
+
+{notes section}
+```
+
+### Skill Template
+
+```markdown
+---
+name: {skill-name}
+description: {description from Claude skill - explain when to trigger}
+---
+
+# {Skill Title}
+
+## Instructions
+
+1. {step 1}
+2. {step 2}
+3. {step 3}
+
+## Guidelines
+
+- {guideline 1}
+- {guideline 2}
+```
+
+### Droid Template
+
+```markdown
+---
+name: {droid-name}
+description: {description of what this droid does}
+model: inherit
+reasoningEffort: medium
+tools: ["Read", "LS", "Grep", "Glob"]
+---
+
+You are a specialized {role} droid.
+
+## Focus
+
+- {behavior 1}
+- {behavior 2}
+- {behavior 3}
+
+## Guidelines
+
+- {guideline 1}
+- {guideline 2}
+```
+
+---
+
+## Step 4: Write Files
+
+**Tools:**
+- `mkdir -p .factory/commands`
+- `mkdir -p .factory/skills/{name}`
+- `mkdir -p .factory/droids`
+- Write(file_path=".factory/commands/{name}.md")
+- Write(file_path=".factory/skills/{name}/SKILL.md")
+- Write(file_path=".factory/droids/{name}.md") - only if custom droid needed
+
+**Skill name mapping:**
+- `.claude/skills/design/figma-extraction/` → `.factory/skills/figma-extraction/`
+- `.claude/skills/architecture/quality-code-check/` → `.factory/skills/quality-code-check/`
+- `.claude/skills/ux/accessibility/` → `.factory/skills/ux-accessibility/`
+
+---
+
+## Step 5: Output Classification
+
+```
+Factory Droid Commands (.factory/commands/):
+  - create-plan.md: [STATUS]
+  - execute-plan.md: [STATUS]
+  - sync-workflow.md: [STATUS]
+
+Factory Droid Skills (.factory/skills/):
+  - figma-extraction/SKILL.md: [STATUS]
+  - quality-code-check/SKILL.md: [STATUS]
+  - frontend-design-fundamentals/SKILL.md: [STATUS]
+
+Factory Droid Droids (.factory/droids/):
+  - code-reviewer.md: [STATUS] (if custom review workflow)
+  - No custom droids needed otherwise
+```
+
+---
+
+## Factory Droid-Specific Notes
+
+1. **Skills are native**: Factory Droid reads skills from `.factory/skills/*/SKILL.md`
+
+2. **AGENTS.md is native**: Factory Droid reads `AGENTS.md` from root - no conversion needed
+
+3. **Command management:**
+   - Reload commands: Press `R` in `/commands`
+   - Import existing commands: Press `I` in `/commands`
+
+4. **Droid tool arrays:**
+   - Array format: `["Read", "LS", "Grep", "Glob"]`
+   - Category string: `"read-only"`, `"edit"`, `"execute"`, `"web"`, `"mcp"`
+   - Combine as needed for custom permissions
+
+5. **Example droids:**
+   - `code-reviewer`: `tools: ["Read", "LS", "Grep", "Glob"]` - read-only review
+   - `security-sweeper`: `tools: ["Read", "Grep", "WebSearch"]` - security analysis
+   - `docs-writer`: `tools: ["Read", "Write", "Edit", "Glob"]` - documentation
+
+6. **Reasoning effort:**
+   - `low`: Fast, simple tasks
+   - `medium`: Balanced (default)
+   - `high`: Complex analysis tasks

package/docs/sync-workflow/github-copilot.md

@@ -0,0 +1,137 @@
+# GitHub Copilot Sync Guide
+
+Sync Claude Code workflows to GitHub Copilot format.
+
+---
+
+## Prerequisites
+
+- Source: `.claude/commands/*.md`, `.claude/CLAUDE.md`
+- Target: `.github/prompts/*.prompt.md`, `.github/copilot-instructions.md`
+
+---
+
+## Step 1: Fetch GitHub Copilot Documentation
+
+**Search queries (try in order until successful):**
+1. `"github.com" copilot custom instructions repository documentation`
+2. `site:docs.github.com copilot custom instructions prompts`
+3. `github copilot prompt files .github/prompts format`
+
+**Key documentation URLs:**
+- `https://docs.github.com/en/copilot/customizing-copilot/adding-repository-custom-instructions-for-github-copilot`
+- `https://docs.github.com/en/copilot/how-tos/configure-custom-instructions`
+
+**Extract and note:**
+- Prompt file location: `.github/prompts/`
+- File format: `*.prompt.md` naming convention
+- Frontmatter structure (applyTo, excludeAgent, etc.)
+- Custom instructions file: `.github/copilot-instructions.md`
+- AGENTS.md support (location, format)
+
+---
+
+## Step 2: Conversion Rules
+
+### Commands → Prompts (`.github/prompts/*.prompt.md`)
+
+**Filename:** `{name}.prompt.md` format
+
+**Frontmatter:**
+- May need `applyTo` for path-specific prompts
+- Keep `name` and `description`
+
+**Tool references conversion:**
+| Claude Code | GitHub Copilot |
+|-------------|----------------|
+| `AskUserQuestion(...)` | Ask user for clarification |
+| `Task(subagent_type='Explore')` | Search codebase for... |
+| `Read(file_path=...)` | Read file |
+| `Write(file_path=...)` | Write file |
+| `Edit(file_path=...)` | Edit file |
+| `Glob(pattern=...)` | Find files |
+| `Grep(pattern=...)` | Search pattern |
+| `Bash(command=...)` | Run command |
+
+**Remove:** Claude-specific skill references (Copilot doesn't have skills system)
+
+**Keep:** Goal, Steps structure, acceptance criteria
+
+### Skills
+
+GitHub Copilot doesn't have a skills system. Options:
+1. Inline skill content into relevant prompts
+2. Add to `copilot-instructions.md` as guidelines
+3. Skip skills (not applicable)
+
+---
+
+## Step 3: Templates
+
+### Prompt Template
+
+```markdown
+---
+name: {command-name}
+description: {description}
+---
+
+## Goal
+
+{goal from Claude command}
+
+## Workflow Alignment
+
+{workflow section - simplified}
+
+## Step 1: {step name}
+
+{converted step content - generic instructions}
+
+...
+
+## Notes
+
+{notes section - remove Claude-specific references}
+```
+
+### copilot-instructions.md Template
+
+```markdown
+# Repository Instructions for GitHub Copilot
+
+## Coding Philosophy
+
+{core philosophy from .claude/CLAUDE.md}
+
+## Workflow Guidelines
+
+{workflow guidelines - converted to generic instructions}
+
+## Code Standards
+
+{code presentation, TODO policy, etc.}
+```
+
+---
+
+## Step 4: Write Files
+
+**Tools:**
+- `mkdir -p .github/prompts`
+- Write(file_path=".github/prompts/{name}.prompt.md")
+- Write(file_path=".github/copilot-instructions.md")
+
+---
+
+## Step 5: Output Classification
+
+```
+GitHub Copilot (.github/prompts/):
+  - create-plan.prompt.md: [STATUS]
+  - execute-plan.prompt.md: [STATUS]
+  - sync-workflow.prompt.md: [STATUS]
+
+GitHub Copilot Base Instructions:
+  - .github/copilot-instructions.md: [STATUS]
+```

package/docs/sync-workflow/opencode.md

@@ -0,0 +1,227 @@
+# OpenCode Sync Guide
+
+Sync Claude Code workflows to OpenCode format.
+
+---
+
+## Prerequisites
+
+- Source: `.claude/commands/*.md`, `.claude/skills/**/SKILL.md`, `.claude/CLAUDE.md`
+- Target: `.opencode/command/*.md`, `.opencode/skill/*/SKILL.md`, `.opencode/agent/*.md`, `AGENTS.md`
+
+---
+
+## Step 1: Fetch OpenCode Documentation
+
+**Search queries (try in order until successful):**
+1. `site:opencode.ai docs agents commands skills`
+2. `opencode.ai documentation configuration`
+3. `opencode ai terminal commands skills agents`
+
+**Key documentation URLs:**
+- `https://opencode.ai/docs/agents/` - Agents configuration
+- `https://opencode.ai/docs/commands/` - Custom commands
+- `https://opencode.ai/docs/skills/` - Agent skills
+- `https://opencode.ai/docs/rules/` - Rules (AGENTS.md)
+- `https://opencode.ai/docs/config/` - Configuration
+
+---
+
+## Step 2: Conversion Rules
+
+### Commands (`.opencode/command/*.md`)
+
+**Location**: `.opencode/command/{name}.md`
+
+**Frontmatter:**
+- `description` (required)
+- `agent: build` for commands that modify code
+- `agent: plan` for read-only commands
+- `model`: optional model override
+- `subtask`: optional subtask flag
+
+**Arguments:**
+- `$ARGUMENTS` for all text after command
+- `$1`, `$2`, etc. for positional args
+
+**Tool references conversion:**
+| Claude Code | OpenCode |
+|-------------|----------|
+| `AskUserQuestion(...)` | (handled by TUI) |
+| `Task(subagent_type='Explore')` | `@explore` mention |
+| `Task(subagent_type='General')` | `@general` mention |
+| `Read(file_path=...)` | `@filename` reference |
+| `Write(file_path=...)` | Write instruction |
+| `Edit(file_path=...)` | Edit instruction |
+| `Bash(command=...)` | `` !`command` `` syntax |
+
+### Skills (`.opencode/skill/<name>/SKILL.md`)
+
+**Location**: `.opencode/skill/{skill-name}/SKILL.md`
+
+**Name validation:**
+- Lowercase alphanumeric with hyphens
+- 1-64 characters
+- Directory name must match `name` in frontmatter
+
+**Frontmatter (required):**
+- `name`: skill name
+- `description`: 1-1024 chars
+- `license`: optional (e.g., MIT)
+- `compatibility`: optional (e.g., opencode)
+- `metadata`: optional key-value pairs
+
+### Agents (`.opencode/agent/*.md`)
+
+**Built-in agents (don't need to create):**
+- `build` - Primary agent with all tools (default)
+- `plan` - Read-only analysis agent
+- `general` - Subagent for complex tasks
+- `explore` - Subagent for codebase exploration
+
+**Create custom agents only if needed.**
+
+**Frontmatter:**
+- `description`: agent description
+- `mode`: `primary` or `subagent`
+- `model`: model identifier
+- `temperature`: optional
+- `tools`: tool permissions object
+- `permission`: skill permissions
+
+---
+
+## Step 3: Templates
+
+### Command Template
+
+```markdown
+---
+description: {description}
+agent: build
+---
+
+{goal and instructions from Claude command}
+
+## Instructions
+
+{step-by-step instructions}
+
+### Step 1: {step name}
+
+{converted step content}
+- Use @explore to search codebase
+- Use @filename to reference files
+- Use !`command` for shell output
+
+...
+
+## Notes
+
+{notes section}
+```
+
+### Skill Template
+
+```markdown
+---
+name: {skill-name}
+description: {description from Claude skill, 1-1024 chars}
+license: MIT
+compatibility: opencode
+metadata:
+  category: {category}
+  source: claude-code
+---
+
+## What I do
+
+{capabilities from Claude skill}
+
+## When to use me
+
+{triggers and conditions from Claude skill}
+
+## Instructions
+
+{step-by-step instructions from Claude skill}
+```
+
+### Agent Template (if custom needed)
+
+```markdown
+---
+description: {agent description}
+mode: subagent
+model: anthropic/claude-sonnet-4-20250514
+temperature: 0.1
+tools:
+  write: false
+  edit: false
+  bash: false
+permission:
+  skill:
+    "*": allow
+---
+
+{system prompt for agent behavior}
+
+Focus on:
+- {behavior 1}
+- {behavior 2}
+- {behavior 3}
+```
+
+---
+
+## Step 4: Write Files
+
+**Tools:**
+- `mkdir -p .opencode/command`
+- `mkdir -p .opencode/skill/{name}`
+- `mkdir -p .opencode/agent`
+- Write(file_path=".opencode/command/{name}.md")
+- Write(file_path=".opencode/skill/{name}/SKILL.md")
+- Write(file_path=".opencode/agent/{name}.md") - only if custom agent needed
+
+**Skill name mapping:**
+- `.claude/skills/design/figma-extraction/` → `.opencode/skill/figma-extraction/`
+- `.claude/skills/architecture/quality-code-check/` → `.opencode/skill/quality-code-check/`
+- `.claude/skills/ux/accessibility/` → `.opencode/skill/ux-accessibility/`
+
+---
+
+## Step 5: Output Classification
+
+```
+OpenCode Commands (.opencode/command/):
+  - create-plan.md: [STATUS]
+  - execute-plan.md: [STATUS]
+  - sync-workflow.md: [STATUS]
+
+OpenCode Skills (.opencode/skill/):
+  - figma-extraction/SKILL.md: [STATUS]
+  - quality-code-check/SKILL.md: [STATUS]
+  - frontend-design-fundamentals/SKILL.md: [STATUS]
+
+OpenCode Agents (.opencode/agent/):
+  - Using built-in agents (build, plan, explore, general)
+  - [custom agent name]: [STATUS] (if any)
+```
+
+---
+
+## OpenCode-Specific Notes
+
+1. **Skills are native**: OpenCode reads `.claude/skills/*/SKILL.md` directly! But for clarity, also create `.opencode/skill/*/SKILL.md`
+
+2. **AGENTS.md is native**: OpenCode reads `AGENTS.md` and `CLAUDE.md` from root - no conversion needed
+
+3. **Command syntax special features:**
+   - Arguments: `$ARGUMENTS` for all, `$1`, `$2` for positional
+   - Shell output: `` !`git status` `` injects command output
+   - File references: `@src/file.ts` includes file content
+
+4. **Agent modes:**
+   - `primary`: Main agents (Tab to switch)
+   - `subagent`: Invoked by `@mention` or by primary agents

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ai-workflow-init",
-  "version": "6.2.3",
+  "version": "6.3.2",
   "description": "Initialize AI workflow docs & commands into any repo with one command",
   "bin": {
     "ai-workflow-init": "./cli.js"