feat(commands): Restructure commands into context/<type>/<action> hierarchy

### Added
- 35 new/moved context commands organized by type:
  - plan: create, review, refine
  - agent: create, review, refine, promote
  - command: create, review, refine, promote, search
  - skill: create, review, refine, promote, search
  - plugin: create, review, refine, brainstorm, issues, research, roadmap, scaffold, search
  - rule: create, review, refine, promote
  - mcp: create, review, refine, promote, search

### Changed
- Moved existing commands to new hierarchy with git mv
- Commands now follow context:<type>:<action> naming

### Notes
- P3 brainstorm commands skipped (low value vs create commands)
- plugin/promote and plugin/plan skipped (plugin workflow is already comprehensive)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: aRustyDev

.claude/commands/create-agent.md .claude/commands/context/agent/create.md.claude/commands/create-agent.md renamed to .claude/commands/context/agent/create.md

@@ -99,9 +99,11 @@ mkdir -p "$(dirname "<target-path>")"
 ## Workflow
 
 ### Step 1: <Phase>
+
 <Instructions>
 
 ### Step 2: <Phase>
+
 <Instructions>
 
 ## Model
@@ -131,9 +133,11 @@ mkdir -p "$(dirname "<target-path>")"
 ## Usage
 
 ### Full Pipeline
+
 <Example invocations>
 
 ### Individual Sub-Agents
+
 <How to run specific stages>
 
 ## Sub-Agents
@@ -152,14 +156,17 @@ mkdir -p "$(dirname "<target-path>")"
 <Detailed instructions>
 
 ### <Sub-Agent 2 Name>
+
 ...
 
 ## Pipeline
 
 ```
+
 Stage 1: <name> → Stage 2: <name> → Stage 3: <name>
                                    ↘ Stage 3b: <name> (parallel)
-```
+
+```text
 
 ## Configuration
 
@@ -183,7 +190,7 @@ Stage 1: <name> → Stage 2: <name> → Stage 3: <name>
 
 ### Step 6: Report
 
-```
+```text
 ## Agent Created
 
 | Field | Value |
@@ -201,7 +208,7 @@ Stage 1: <name> → Stage 2: <name> → Stage 3: <name>
 
 ## Examples
 
-```
+```text
 /create-agent homebrew-expert --location context
 /create-agent code-reviewer --type solo
 /create-agent skill-reviewer --type orchestrated --location project
@@ -257,15 +264,18 @@ For orchestrated agents, assign the cheapest model that can handle each sub-agen
 ## Troubleshooting
 
 **Agent not being invoked:**
+
 - Ensure the file is in `.claude/agents/` or `context/agents/`
 - Check the file has `.md` extension
 - Verify the first line is a `# Title` heading
 
 **Sub-agents not executing in parallel:**
+
 - Task tool calls must be in the same message to run in parallel
 - Verify sub-agents have no data dependencies between them
 
 **Agent producing inconsistent results:**
+
 - Add explicit validation steps after each phase
 - Include example inputs/outputs in the agent definition
 - Constrain the tool list to only what's needed

.claude/commands/context/agent/promote.md

@@ -0,0 +1,124 @@
+---
+description: Promote an agent to the AI Config Library
+argument-hint: <agent-path>
+allowed-tools: Read, Write, Bash(git:*), Bash(gh:*), Bash(mkdir:*), Bash(cp:*), AskUserQuestion
+---
+
+# Promote Agent
+
+Promote an agent from the current project to the aRustyDev/ai repository.
+
+## Arguments
+
+- `$1` - Path to the agent file. Example: `.claude/agents/terraform-planner.md`
+
+## Configuration
+
+The ai repo location is determined in order of precedence:
+
+1. Environment variable `$AI_CONFIG_REPO` if set
+2. Git submodule named `ai` in current repo's parent
+3. Default: `~/repos/configs/ai`
+
+```bash
+AI_REPO="${AI_CONFIG_REPO:-$(git config --file .gitmodules --get submodule.ai.path 2>/dev/null || echo "$HOME/repos/configs/ai")}"
+```
+
+## Workflow
+
+### Phase 1: Validate and Analyze
+
+1. **Verify agent exists**:
+   - Confirm the agent file exists at the provided path
+   - Verify proper markdown structure with frontmatter
+
+2. **Extract agent metadata**:
+   - Parse frontmatter for name/description
+   - Identify any supporting files
+   - Note tool requirements
+
+3. **Check for existing agent in ai repo**:
+
+   ```bash
+   ls "$AI_REPO/context/agents/"
+   ```
+
+   - If found, compare and determine: Identical, Extension, Upgrade, or Separate
+
+4. **Check for existing branch/issue/PR**:
+
+   ```bash
+   git -C "$AI_REPO" branch --list "feat/add-agent-<name>"
+   gh issue list --repo aRustyDev/ai --search "[AGENT] <name> in:title"
+   gh pr list --repo aRustyDev/ai --search "feat(agent): add <name> in:title"
+   ```
+
+### Phase 2: User Decision (if existing)
+
+If existing agent found, ask user:
+
+1. **Upgrade existing** - Replace with new version
+2. **Extend existing** - Merge new content
+3. **Create separate** - Use different name
+4. **Cancel** - Abort
+
+### Phase 3: Create Issue
+
+Create GitHub Issue with:
+
+- **Agent Name**: From frontmatter
+- **Description**: From frontmatter
+- **Source Path**: Original path
+- **Agent Content**: Summary or embedded
+- **Tool Requirements**: List from frontmatter
+
+### Phase 4: Create Feature Branch and PR
+
+1. **Create feature branch**:
+
+   ```bash
+   git -C "$AI_REPO" checkout main && git -C "$AI_REPO" pull
+   git -C "$AI_REPO" checkout -b feat/add-agent-<name>
+   ```
+
+2. **Copy agent file**:
+
+   ```bash
+   cp "<path>" "$AI_REPO/context/agents/<name>.md"
+   ```
+
+3. **Commit and push**:
+
+   ```bash
+   git -C "$AI_REPO" add context/agents/<name>.md
+   git -C "$AI_REPO" commit -m "feat(agent): add <name>"
+   git -C "$AI_REPO" push -u origin feat/add-agent-<name>
+   ```
+
+4. **Create PR**:
+
+   ```bash
+   gh pr create --repo aRustyDev/ai \
+     --head feat/add-agent-<name> --base main \
+     --title "feat(agent): add <name>" \
+     --body "..."
+   ```
+
+### Phase 5: Report
+
+| Item | URL |
+|------|-----|
+| Issue | `<url>` |
+| PR | `<url>` |
+
+## Examples
+
+```bash
+/context:agent:promote .claude/agents/terraform-planner.md
+/context:agent:promote context/agents/code-reviewer.md
+```
+
+## Related Commands
+
+- `/context:agent:create` - Create a new agent
+- `/context:agent:review` - Review agent quality

.claude/commands/context/agent/refine.md

@@ -0,0 +1,73 @@
+---
+description: Refine an agent definition based on review feedback
+argument-hint: <agent-path>
+allowed-tools: Read, Write, Edit, AskUserQuestion
+---
+
+# Refine Agent
+
+Improve an agent definition based on review findings or user feedback.
+
+## Arguments
+
+- `$1` - Path to agent file (required)
+
+## Workflow
+
+### Step 1: Load and Analyze
+
+1. Read agent file
+2. Run quick review to identify issues
+3. Present findings to user
+
+### Step 2: Gather Feedback
+
+Use AskUserQuestion:
+
+1. Which issues to address?
+2. Any additional changes needed?
+
+### Step 3: Apply Refinements
+
+Common refinements:
+
+- Add missing sections
+- Adjust model assignments
+- Reduce tool scope
+- Clarify workflow steps
+- Add error handling
+- Include examples
+
+### Step 4: Validate
+
+1. Re-run review checks
+2. Verify structure is valid
+3. Confirm no regressions
+
+### Step 5: Report
+
+```text
+## Agent Refined
+
+| Field | Value |
+|-------|-------|
+| Agent | <name> |
+| Changes | N |
+
+**Modifications:**
+- <change 1>
+- <change 2>
+
+Run `/context:agent:review` to verify.
+```
+
+## Examples
+
+```bash
+/context:agent:refine .claude/agents/code-reviewer.md
+```
+
+## Related Commands
+
+- `/context:agent:review` - Review agent quality
+- `/context:agent:create` - Create new agent