Command creator

Command Creator

Create Claude Code slash commands — reusable workflows invoked with /command-name. A slash command is a markdown file in .claude/commands/ (project-level) or ~/.claude/commands/ (global) that expands into a prompt when invoked. Good candidates: repetitive workflows (code review, PR submission, CI fixing), multi-step processes that need consistency, agent delegation patterns, and project-specific automation.

Bundled Resources

Load these as needed during the workflow below:

• references/patterns.md — command patterns (workflow automation, iterative fixing, agent delegation, simple execution)
• references/examples.md — real command examples with full source (submit-stack, ensure-ci, create-implementation-plan)
• references/best-practices.md — template structure, writing guidelines, quality checklist, common pitfalls

Workflow

Step 1: Determine Location

Check git rev-parse --is-inside-work-tree 2>/dev/null: inside a git repo → project-level .claude/commands/; otherwise → global ~/.claude/commands/. An explicit user request ("global", "user-level", "project") overrides the default. Report the chosen location before proceeding.

Step 2: Identify Command Pattern

Load references/patterns.md to understand the available patterns:

• Workflow Automation — Analyze → Act → Report (e.g., submit-stack)
• Iterative Fixing — Run → Parse → Fix → Repeat (e.g., ensure-ci)
• Agent Delegation — Context → Delegate → Iterate (e.g., create-implementation-plan)
• Simple Execution — Run command with args (e.g., codex-review)

Infer the best pattern from the user's description. Only ask if genuinely ambiguous.

Step 3: Gather Command Information

Interview the user for whatever isn't already clear from their request:

• Name and purpose — kebab-case filename, matching the invocation (my-command.md → /my-command); a concise, action-oriented description for /help output.
• Arguments — required or optional, and what they represent. If the command takes arguments, add argument-hint: to the frontmatter (<angle-brackets> for required, [square-brackets] for optional).
• Workflow steps — the order of actions, tools/commands to use, how to handle results, success criteria, and error handling.
• Tool guidance — specific agents/tools to use or avoid, files to read for context.

Step 4: Generate the Command

Load references/best-practices.md for the template structure, writing style guidelines, and quality checklist. Key principles: imperative verb-first instructions, explicit and specific actions, stated expected outcomes, concrete examples, clear error handling.

Step 5: Create the File

Ensure the target directory exists, write the command file, then confirm with the user: report the file location, summarize what the command does, and show how to invoke it (/command-name [arguments]).

Step 6: Test and Iterate (Optional)

Suggest testing with /command-name [arguments] and iterate on feedback.