Agent Skills: command-creator

  <substep n="1a" title="Track Progress">
    <action if="first iteration">Create todo list from all instruction step titles (TodoWrite)</action>
    <action>Mark current step as in_progress (TodoWrite)</action>
  </substep>

  <substep n="1b" title="Handle Step Attributes">
    <check>If optional="true" and NOT YOLO -> Use Question tool to ask user whether to include (Yes/Skip)</check>
    <check>If if="condition" -> Evaluate condition</check>
    <check>If for-each="item" -> Repeat step for each item</check>
    <check>If repeat="n" -> Repeat step n times</check>
  </substep>

  <substep n="1c" title="Execute Step Content">
    <action>Process step instructions (markdown or XML tags)</action>
    <action>Replace {{variables}} with values (ask user if unknown)</action>
    <execute-tags>
      <tag>action -> Perform the action</tag>
      <tag>action type="script" -> Execute as shell command, capture output</tag>
      <tag>check if="condition" -> Conditional block wrapping actions (requires closing &lt;/check&gt;)</tag>
      <tag>ask -> Prompt user using Question tool and WAIT for response. When options exist, present them as selectable choices. For open-ended questions, provide helpful suggested options with custom input enabled</tag>
      <tag>invoke-skill name="skill-name" -> Execute another skill via OpenCode skill system</tag>
      <tag>invoke-agent name="subagent" prompt="task" -> Delegate to subagent via Task tool</tag>
      <tag>goto step="x" -> Jump to specified step (used for explicit loop-back in Q-R style flows)</tag>
    </execute-tags>
  </substep>

  <substep n="1d" title="Handle template-output Tags">
    <if tag="template-output">
      <mandate>Generate content for this section</mandate>
      <mandate>Save to file (Write first time, Edit subsequent)</mandate>
      <action>Display generated content</action>
      <action>Use Question tool to present checkpoint options. WAIT for response.
        <options>
          <option label="Continue" description="Approve this section and move to the next step" />
          <option label="Advanced Elicitation" description="Explore this section deeper before proceeding" />
          <option label="YOLO" description="Auto-complete the rest of this workflow without stopping" />
        </options>
        <if response="Advanced Elicitation">
          <action>Invoke advanced-elicitation skill</action>
        </if>
        <if response="Continue">
          <action>Continue to next step</action>
        </if>
        <if response="YOLO">
          <action>Enter YOLO mode for the rest of the workflow</action>
        </if>
      </action>
    </if>
  </substep>

  <substep n="1e" title="Step Completion">
    <action>Mark current step as completed (TodoWrite)</action>
    <check>If no special tags and NOT YOLO:</check>
    <action>Use Question tool to ask:
      <options>
        <option label="Continue" description="Proceed to the next step" />
        <option label="Edit" description="Revise this step before moving on" />
      </options>
    </action>
  </substep>
</step>

<step n="2" title="Completion">
  <check>Confirm all instruction steps have been executed</check>
  <output>Workflow complete.</output>
</step>

<output>
  A complete command directory at .vera/commands/{command_name}/
  containing the command source files. After build, the command appears
  under the configured target (e.g., ~/.config/opencode/command/).
</output>

<key-responsibility>
  The critical decision is COMMAND TYPE:

  1. STANDALONE — Self-contained instructions with optional workflow.
     Example: /commit, /github — complete instructions for a task.

  2. WRAPPER — Thin shell that delegates to a skill.
     Example: /pr-review — just invokes the pr-review skill.

  3. AGENT-SPECIFIC — Lives in agent subfolder, only available to that
     agent. Example: /prompter/verify-compliance — only for Percy.

  Commands are intentionally simpler than skills. Reach for skill-creator
  instead when the work needs multi-step state, branching logic, or
  sub-skill composition.
</key-responsibility>

<step n="1" title="Discovery and Context">
  <action>Load and execute: references/steps/step-01-discovery.xml</action>
  <note>
    Gather requirements: what the command does, when it's triggered,
    and what arguments it accepts. Confirm a command is the right
    primitive — if the work needs state machine, branching, or
    sub-skill composition, redirect to skill-creator.
  </note>
</step>

<step n="2" title="Command Type" critical="true">
  <action>Load and execute: references/steps/step-02-type.xml</action>
  <note>
    THIS IS THE KEY DECISION. Determines:
    - Standalone vs Wrapper vs Agent-specific
    - Whether it needs a workflow structure
    - Whether it delegates to a skill
  </note>
</step>

<step n="3" title="Scaffold Command">
  <action>Load and execute: references/steps/step-03-scaffold.xml</action>
  <note>
    Create the command directory and initial files from template.
    Commands output single .md files but source has config +
    instructions.
  </note>
</step>

<step n="4" title="Command Identity">
  <action>Load and execute: references/steps/step-04-identity.xml</action>
  <note>
    Define name, description, agent assignment, and argument-hint. The
    description is critical for discoverability.
  </note>
</step>

<step n="5" title="Create Instructions">
  <action>Load and execute: references/steps/step-05-instructions.xml</action>
  <note>
    Write the command body — the actual instructions. Standalone
    commands may have workflows; wrappers are minimal.
  </note>
</step>

<step n="6" title="Validate and Complete">
  <action>Load and execute: references/steps/step-06-validate.xml</action>
  <note>
    Run validation on the completed files. Update vera.manifest.yaml if
    the command joins an existing build group.
  </note>
</step>

<next-steps>
  <step>Build dubstack: `vera build`</step>
  <step>Test by running: /{command_name} [args]</step>
  <step>If agent-specific, ensure the agent has access via its permission map</step>
</next-steps>