From drupal-workflow
Dispatches specialized Drupal agents in parallel to execute implementation plans with independent tasks. Use after planning for multi-file or multi-module changes.
npx claudepluginhub gkastanis/drupal-workflow --plugin drupal-workflowThis skill uses the workspace's default tool permissions.
Parallel execution of implementation plans using the plugin's specialized agents.
Creates detailed implementation plans for sub-agent delegation or engineers lacking codebase context, with exact file paths, complete code, verification commands, and bite-sized tasks.
Dispatches subagents via Claude Code's Agent tool to parallelize 3+ independent tasks across different files/modules with verification commands. Activates on execution-ready plans or keywords like 'use subagents'.
Orchestrates implementation plan execution by grouping tasks into subsystems, spawning sub-agents for parallel processing of independent groups and sequential dependencies.
Share bugs, ideas, or general feedback.
Parallel execution of implementation plans using the plugin's specialized agents.
drupal-brainstorming or writing-plans produced a planBefore dispatching agents:
drupal-brainstorming first/drupal-status to check. Agents need accurate structural dataUse the right agent for each task type:
| Task Type | Agent | Model | When |
|---|---|---|---|
| New code, services, entities, hooks | @drupal-builder | opus | Implementation work |
| Architecture review, security audit | @drupal-reviewer | sonnet | After implementation |
| Smoke tests, drush eval, config checks | @drupal-verifier | sonnet | After implementation |
| Semantic docs, tech specs | @semantic-architect | sonnet | After structural changes |
Dispatch in parallel when tasks are independent:
Agent({ subagent_type: "drupal-workflow:drupal-builder", description: "Implement X", prompt: "..." })
Agent({ subagent_type: "drupal-workflow:drupal-builder", description: "Implement Y", prompt: "..." })
Dispatch sequentially when one depends on another:
1. @drupal-builder implements the feature
2. @drupal-reviewer reviews the implementation
3. @drupal-verifier verifies it works
Each agent prompt MUST be self-contained. The agent has NO context from this conversation.
Include in every prompt:
Bad prompt: "Implement the feature from the plan" Good prompt: "Create a service at code/my_module/src/Service/Calculator.php that implements CalculatorInterface with method add(int, int): int. Register in my_module.services.yml with interface type-hint. Verify with: ddev drush eval '...'"
Create a task for each delegated unit:
TaskCreate({ description: "Implement Calculator service" })
TaskCreate({ description: "Implement admin form" })
TaskCreate({ description: "Verify all services" })
Update tasks as agents complete. This creates a visible progress trail.
After all agents finish:
discover deps:FEATURE to confirm no unintended side effectsrun_in_background: true for agents that don't block subsequent workgeneral-purpose agents for Drupal work (use specialized agents)drupal-brainstorming → writing-plans → drupal-delegation → verification-before-completion
↑ ↓
└─── If delegation reveals gaps ───────┘