drupal-workflow

Analyze dependencies and blast radius for a feature or module

Auto-detect project state and bootstrap structural index, semantic docs, and CLAUDE.md hint

Load full project context into session (debug/overview — heavy, ~2500 tokens)

Regenerate structural index and update CLAUDE.md hint

Generate and manage semantic documentation (business index, tech specs, business schemas)

Check documentation, structural index, and staleness status

Run Drupal test script to verify implementation

Verify Drupal implementation using smoke tests and drush checks

Implement changes across all affected files with validation

Verify code changes are complete and consistent

**Role**: Full-stack Drupal implementation -- modules, themes, config, migrations, performance.

**Role**: Architecture decisions, security audit, coding standards validation, and test writing.

**Role**: Implementation verification via `ddev drush eval`, curl smoke tests, and config checks. Read-only -- verifies, does not fix.

**Role**: Read the structural index (Layer 2) + source code to produce semantic documentation (Layer 3) — business index, tech specs, and business schemas.

Explore requirements, constraints, and design options before implementing Drupal features. Use BEFORE writing code — when starting a new feature, module, or significant change. Produces a decision record with entity design, service architecture, and hook/event strategy.

Cache bins, tags, contexts, invalidation strategies, and external caching for Drupal 10/11. Use when implementing caching, optimizing performance, or configuring cache backends.

PHPCS, PHPStan, naming conventions, and code style enforcement for Drupal 10/11. Use when checking coding standards, running static analysis, or enforcing code quality.

Configuration management patterns for Drupal 10/11 — config split, config ignore, environments, import/export workflows, config readonly, and the sync/install/optional directory hierarchy. Use when working with config export/import, multi-environment deployments, config split setup, or managing configuration across dev/staging/prod.

Load Drupal-specific conventions (translations, CSS, error handling) on demand. Use when working on theming, translations, or exception handling.

Execute implementation plans by dispatching specialized Drupal agents in parallel. Use AFTER planning — when you have a plan with independent tasks to delegate. Dispatches @drupal-builder, @drupal-reviewer, @drupal-verifier with structural awareness.

Field type selection, entity CRUD operations, view modes, and content modeling patterns for Drupal 10/11. Use when designing content types, selecting field types, or working with the Entity API.

OOP hooks (Drupal 11), form alters, entity hooks, and legacy bridges for Drupal 10/11. Use when implementing hooks, form alterations, or event subscribers.

Core Drupal development rules covering code quality, security, services, and testing verification. ALWAYS consult when writing Drupal code.

OWASP prevention patterns, access control, input sanitization, and XSS protection for Drupal 10/11. Use when reviewing code for security issues, implementing access control, or hardening Drupal applications.

Service definitions, dependency injection patterns, and interface design for Drupal 10/11. Use when creating services, registering dependencies, or implementing the service container pattern.

Practical Drupal testing patterns using curl smoke tests, drush eval, and test scripts. Use when verifying Drupal implementations actually work - module enabled, service exists, page renders, config correct.

Analyze autopilot effectiveness and tune the drupal-workflow plugin based on real session data. Use when the autopilot feels too noisy, too quiet, or is being ignored. Also use when you want to check if recent changes to the plugin are actually working, review intervention acceptance rates, or improve classifier accuracy. Trigger on: "tune autopilot", "autopilot stats", "self-improve", "is the autopilot working", "too many warnings", "interventions being ignored", "plugin health", or after any conversation about autopilot effectiveness.

Docs-first discovery for efficient codebase exploration. Use before Glob/Grep/Explore to get Logic IDs and file paths directly from semantic documentation. Saves tokens by avoiding expensive file searches. Commands include `/discover FEATURE` for feature lookup, `/discover "search terms"` for keyword search, and `/discover --prime` to load the business index.

Navigate semantic documentation in docs/semantic/ that maps business logic to Drupal code. Use this skill when asking how features work (authentication, content types, views, migrations, access control, workflows, caching, theming, forms, blocks, entities). Search by Logic ID (AUTH-L1), feature code (MIGR), user story (US-001), or entity schema. Answers "where is X implemented?", "how does X work?", "what code handles X?".

Generate and query structural indexes for Drupal projects. Parses services.yml, routing.yml, hooks, plugins, and entity types to build an automated dependency graph. Use for blast radius analysis, service dependency mapping, and structural queries. Commands include `/structural-index` to generate, and structural query prefixes via discover (e.g., `discover service:entity_type.manager`, `discover deps:AUTH`).

Twig template patterns, filters, theme suggestions, and component architecture for Drupal 10/11. Use when creating or modifying Twig templates, implementing theme hooks, or building front-end components.

Gate function ensuring implementations are tested before claiming completion. Prevents rationalization and untested claims. Use at the end of any implementation task to validate work actually functions.

Use when delegating tasks to sub-agents or creating implementation plans. Write comprehensive plans assuming the engineer has zero codebase context. Include exact file paths, complete code, verification commands, and bite-sized tasks.

13 hooks across 6 events