chore: remove command-authoring and command-audit skills and references

Slash commands are no longer maintained — skills handle that work now.

- Delete skills/command-authoring/ (5 files) and skills/command-audit/
  (11 files)
- Remove Skill(command-audit) permission from settings.json
- Remove audit-command auditor and Command entries from
  audit-coordinator decision matrix and component type lists
- Remove author-command from Related Skills in skill-authoring,
  agent-authoring, and output-style-authoring
- Rewrite agent-decision-guide.md as "Agents vs Skills" (remove
  Command column, section, decision flow branch, migration path)
- Remove command from evaluator type-specific validation list
- Remove Slash Commands section, Command→Skill Reference Table, and
  Skills vs Commands Decision Guide from naming-conventions.md
- Remove Commands section and validation checklist from
  frontmatter-requirements.md
- Remove Command from decision matrix, scenarios, and quick reference
  in when-to-use-what.md
- Remove commands/ directory section and git tracking entries from
  file-organization.md
- Remove Commands section from customization-examples.md
- Clean up stale references in delegation-patterns.md,
  plugin-structure-comparison.md, README.md, references/README.md,
  agent-audit/SKILL.md, and hook-audit/comparison-with-official.md

Co-Authored-By: Claude Opus 4.5 <[email protected]>

Author: philoserf

README.md

@@ -101,23 +101,6 @@ The **skill-authoring** skill guides you through:
 
 **Examples**: Best practices, auditing guidelines, deployment procedures
 
-### Creating Commands
-
-**When to use**: Build quick shortcuts for specific workflows or as entry points to skills.
-
-```bash
-/create-command my-command
-```
-
-The **command-authoring** skill guides you through:
-
-- Designing delegation patterns
-- Handling arguments and validation
-- Keeping commands focused and simple
-- Deciding when to delegate to skills
-
-**Examples**: User shortcuts, project templates, explicit workflows
-
 ### Creating Output-Styles
 
 **When to use**: Define persona modes that change how Claude behaves (tone, verbosity, approach).

references/README.md

@@ -21,14 +21,14 @@ references/
 
 ## Customization Development References
 
-All files in `customization/` support creating and auditing Claude Code extensions (skills, agents, commands, hooks, output-styles).
+All files in `customization/` support creating and auditing Claude Code extensions (skills, agents, hooks, output-styles).
 
 ### Decision Guides
 
 **[customization/decision-matrix.md](customization/decision-matrix.md)** (90 lines)
 
 - Quick reference comparison table
-- Covers 5 component types: Skills, Subagents, Commands, Output Styles, Hooks
+- Covers 4 component types: Skills, Subagents, Output Styles, Hooks
 - Use when: Need fast lookup of component characteristics
 - Links to: when-to-use-what.md for details
 
@@ -76,7 +76,7 @@ All files in `customization/` support creating and auditing Claude Code extensio
   file references
 - Clear vs unclear delegation examples
 - Use when: Writing commands or validating delegation clarity
-- Referenced by: command-audit skill, command-authoring skill
+- Referenced by: authoring and audit skills
 
 **[customization/hook-events.md](customization/hook-events.md)** (241 lines)
 
@@ -177,7 +177,7 @@ These shared files are used by:
 
 - `audit-skill` skill (references all 3 spec files)
 - `agent-authoring` skill (references when-to-use-what.md)
-- `command-authoring` skill (references when-to-use-what.md)
+- `output-style-authoring` skill (references when-to-use-what.md)
 - `output-style-authoring` skill (references when-to-use-what.md)
 
 ## Maintenance

references/customization-examples.md

@@ -35,42 +35,6 @@ Practical examples of well-named and organized Claude Code components.
 └── bug-fixer.md             # Investigates and fixes bugs
 ```
 
-## Commands
-
-### Simple Action Commands
-
-```text
-.claude/commands/
-├── test.md                  # Run tests with optional pattern
-├── build.md                 # Build the project
-├── lint.md                  # Run linters
-└── format.md                # Format code
-```
-
-### Multi-Step Workflow Commands
-
-```text
-.claude/commands/
-├── deploy.md                # Deploy to production
-├── release.md               # Create release (changelog, tag, publish)
-├── security-audit.md        # Run security checks
-└── performance-check.md     # Run performance benchmarks
-```
-
-### Organized by Domain
-
-```text
-.claude/commands/
-├── frontend/
-│   ├── component.md         # /component - Creates React component
-│   ├── page.md              # /page - Creates new page
-│   └── styling.md           # /styling - Generates CSS/styled-components
-└── backend/
-    ├── endpoint.md          # /endpoint - Creates API endpoint
-    ├── migration.md         # /migration - Creates DB migration
-    └── model.md             # /model - Creates data model
-```
-
 ## Skills
 
 ### Reference Skills
@@ -134,9 +98,6 @@ Practical examples of well-named and organized Claude Code components.
 .claude/
 ├── agents/
 │   └── test-runner.md
-├── commands/
-│   ├── test.md
-│   └── test-watch.md
 ├── skills/
 │   └── testing-guide/
 │       ├── SKILL.md
@@ -167,24 +128,6 @@ You are a test specialist. When invoked:
 5. Report results clearly
 ```
 
-### Command: test.md
-
-```yaml
----
-description: Run tests with optional pattern filter
-argument-hint: [pattern]
----
-
-Run the project's test suite.
-
-Usage:
-- `/test` - Run all tests
-- `/test Button` - Run tests matching "Button"
-- `/test integration` - Run integration tests
-
-Delegate to the test-runner agent to analyze failures and suggest fixes.
-```
-
 ### Skill: testing-guide/SKILL.md
 
 ```yaml
@@ -228,6 +171,5 @@ See references/ for framework-specific patterns.
 This creates a complete testing workflow:
 
 - **Agent** for intelligent test fixing
-- **Commands** for quick test execution
 - **Skill** for testing knowledge
 - **Hook** for pre-commit validation

references/delegation-patterns.md

@@ -357,11 +357,9 @@ First read all agent files, then validate each one, then compare with settings.j
 - skill-audit skill
 - agent-audit skill
 - hook-audit skill
-- command-audit skill
 - output-style-audit skill
 - agent-authoring skill
 - skill-authoring skill
-- command-authoring skill
 - output-style-authoring skill
 - git-workflow skill
 - audit-coordinator skill

references/file-organization.md

@@ -11,11 +11,6 @@ Directory structure and layout best practices for Claude Code customizations.
 ├── agents/                    # Specialized AI agents
 │   ├── agent-name.md          # Agent definition
 │   └── another-agent.md
-├── commands/                  # Slash commands
-│   ├── command-name.md        # Command definition
-│   ├── subdirectory/          # Optional grouping
-│   │   └── grouped-cmd.md     # Shows as "project:subdirectory" in /help
-│   └── another-command.md
 ├── skills/                    # Model-invoked capabilities
 │   ├── skill-name/            # Skill directory
 │   │   ├── SKILL.md           # Primary skill definition (required)
@@ -61,22 +56,6 @@ skill-name/
 4. **No Orphans**: All reference files should be discoverable
 5. **Logical Grouping**: Group non-markdown resources by purpose (scripts/, templates/, config/)
 
-## Command Subdirectory Organization
-
-Commands can be organized in subdirectories for grouping:
-
-```text
-.claude/commands/
-├── frontend/
-│   ├── component.md    # Creates /component (shows "project:frontend")
-│   └── styling.md      # Creates /styling (shows "project:frontend")
-└── backend/
-    ├── api.md          # Creates /api (shows "project:backend")
-    └── migration.md    # Creates /migration (shows "project:backend")
-```
-
-**Note**: Subdirectories provide organizational namespacing in `/help` output but do not prefix the command name itself.
-
 ## Agent Organization
 
 Agents with reference materials can optionally use a directory:
@@ -145,7 +124,6 @@ These directories are created automatically and should not be tracked:
 **Always track**:
 
 - `agents/` - Custom agents
-- `commands/` - Slash commands
 - `skills/` - Agent skills
 - `hooks/` - Hook scripts
 - `settings.json` - Configuration (if no secrets)
@@ -171,6 +149,6 @@ These directories are created automatically and should not be tracked:
 5. **Consistent Naming**: Follow kebab-case for all directories and files
 6. **Logical Grouping**: Group related files by purpose (references/, scripts/, etc.)
 7. **No Orphans**: All files should be discoverable and have a clear purpose
-8. **Track Customizations**: Commit agents, commands, skills, and hooks
+8. **Track Customizations**: Commit agents, skills, and hooks
 9. **Ignore Session Data**: Don't commit todos, plans, logs, or history
 10. **Document Structure**: Add README.md files for complex directory structures

references/frontmatter-requirements.md

@@ -73,44 +73,6 @@ allowed_tools:
 ---
 ```
 
-## Commands
-
-```yaml
----
-description: Brief description # Required for /help and model invocation
-argument-hint: [arg1] [arg2] # Optional: shown in autocomplete
-allowed-tools: Bash, Read # Optional: restricts available tools
-model: haiku # Optional: specific model
-disable-model-invocation: false # Optional: prevents SlashCommand tool access
----
-```
-
-### Required Fields
-
-- **description**: Brief command description (required for `/help` visibility and model invocation)
-
-### Optional Fields
-
-- **argument-hint**: Shows expected arguments in autocomplete
-- **allowed-tools**: Restricts which tools the command can use
-- **model**: Override the default model for this command
-- **disable-model-invocation**: Set to `true` to prevent SlashCommand tool from invoking
-
-### Important Notes
-
-- Commands **without** `description` won't appear in `/help`
-- Commands **without** `description` can't be invoked via SlashCommand tool
-- Use `$ARGUMENTS` for all arguments, `$1`, `$2` for positional parameters
-
-### Example
-
-```yaml
----
-description: Validates a sub-agent configuration for correctness, clarity, and effectiveness
-argument-hint: [agent-name]
----
-```
-
 ## Skills
 
 ```yaml
@@ -219,12 +181,6 @@ description: When to use this style
 - [ ] `model` is valid (if specified)
 - [ ] `tools` list matches actual usage (if specified)
 
-### For Commands
-
-- [ ] `description` is present (required for `/help`)
-- [ ] `argument-hint` matches actual usage (if specified)
-- [ ] Variables use `$ARGUMENTS`, `$1`, `$2` format
-
 ### For Skills
 
 - [ ] `name` is kebab-case, lowercase only