- Replace "comprehensive" and other marketing terms with plain language
- Shorten overly detailed HTTP request documentation
- Simplify import structure and data resolution sections
- Remove verbose environment variables explanations
- Cut scheduled workflows examples to essential information
- Make language more direct and honest about development status
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Replace JSONPath references with Handlebars syntax
- Add comprehensive template examples and type conversion docs
- Update CHANGELOG with v0.0.39 breaking changes and migration notes
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Delete unused test files: `basic.test.ts`, `condition-fix.spec.ts`, `create-document-step.test.ts`, and `error-scenarios.spec.ts`
- Streamline codebase by eliminating redundant and outdated test cases
- Improve maintainability by keeping only relevant and up-to-date tests
- Remove cron-trigger.ts and cron-scheduler.ts files
- Clean up cron-related code from plugin index and workflow hooks
- Remove cron references from workflow executor types
- Add cron to NOT-IMPLEMENTING.md with webhook alternatives
- Update README with scheduled workflow documentation using external services
- Suggest GitHub Actions and Vercel Cron as reliable alternatives
Benefits of external scheduling:
- Better reliability and process isolation
- Easier debugging and monitoring
- Leverages existing cloud infrastructure
- Reduces plugin complexity and maintenance burden
- Change default log level to 'warn' for production
- Add PAYLOAD_AUTOMATION_LOG_LEVEL environment variable
- Remove all verbose config-phase logs
- Add documentation for log level control
Major Features:
• Add persistent error tracking for timeout/network failures that bypasses PayloadCMS output limitations
• Implement smart error classification (timeout, DNS, connection, network) with duration-based detection
• Add comprehensive test infrastructure with MongoDB in-memory testing and enhanced mocking
• Fix HTTP request handler error preservation with detailed context storage
• Add independent execution tracking with success/failure status and duration metrics
Technical Improvements:
• Update JSONPath documentation to use correct $.trigger.doc syntax across all step types
• Fix PayloadCMS job execution to use runByID instead of run() for reliable task processing
• Add enhanced HTTP error handling that preserves outputs for 4xx/5xx status codes
• Implement proper nock configuration with undici for Node.js 22 fetch interception
• Add comprehensive unit tests for WorkflowExecutor with mocked PayloadCMS instances
Developer Experience:
• Add detailed error information in workflow context with URL, method, timeout, attempts
• Update README with HTTP error handling patterns and enhanced error tracking examples
• Add test helpers and setup infrastructure for reliable integration testing
• Fix workflow step validation and JSONPath field descriptions
Breaking Changes: None - fully backward compatible
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Create dedicated /server export for server-side functions and plugin
- Main export now contains only types (client-safe)
- Remove problematic /steps export that exposed server functions
- Update README with correct import structure
- Fix tests to use /server export
This prevents server-side code from being bundled in client JavaScript,
eliminating the "require is not defined in ES module scope" runtime error.
Breaking change: workflowsPlugin must now be imported from '/server':
import { workflowsPlugin } from '@xtr-dev/payload-automation/server'
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
