Adds a dummy skill directory that is NOT listed in the README table.
The readme-check workflow should detect this discrepancy and auto-update
the README to include the new skill.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add documentation to prevent common configuration errors:
- entities-create.md: Add warning about nested "schema" property mistake
that causes "Invalid schema: Schema must have a 'type' field" errors
- agents-push.md:
- Fix entity_name example to use PascalCase (Task not tasks)
- Add agent file naming convention (underscores, not hyphens)
- Add warning about wrong "tools" vs "tool_configs" format
- Add best practices for agent instructions when using entity tools
These issues were discovered during real-world usage and caused
significant debugging time.
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Update SDK documentation to recommend building custom login UI using
loginViaEmailPassword() and loginWithProvider() instead of redirectToLogin().
Changes:
- Add deprecation warning to redirectToLogin in method reference table
- Update Protected Route Pattern to use navigate() instead of redirectToLogin
- Update error handling examples to navigate to custom login page
- Reframe "Limitations" section as "Authentication UI Options"
- Add new Best Practice for building custom login UI
- Update QUICK_REFERENCE.md with warning comment
- Update SKILL.md Protected Routes example
https://claude.ai/code/session_01S8pH8AXWza6WkTZ233N8Z8
Co-authored-by: Claude <noreply@anthropic.com>
* Refactor RLS documentation for clarity and structure
- Updated entities-create.md to streamline the Row Level Security (RLS) section, enhancing descriptions and examples for better understanding.
- Consolidated RLS operations and permission values, clarifying the conditions and template variables used in RLS rules.
- Revised examples in rls-examples.md to reflect new patterns and best practices, emphasizing the limitations of JSON schema RLS and alternative approaches for complex access patterns.
- Improved overall organization of content for easier navigation and comprehension.
* Update documentation to clarify AI agent usage requirements in Base44 SDK
- Added a note in SKILL.md indicating that chatting with AI agents requires a logged-in user.
- Included a similar note in base44-agents.md to emphasize that all agent methods operate within the context of an authenticated user.
- Added a "Common Mistakes" section in functions-create.md to guide users on correct directory structure and import statements.
- Updated SKILL.md to include critical warnings against hallucinating API methods, with detailed comparisons of incorrect vs. correct usage.
- Introduced a new QUICK_REFERENCE.md file for quick method signatures across SDK modules, emphasizing the importance of verification before coding.
- Clarified required directory structure for backend functions in functions.md, ensuring users follow the correct setup for deployment.
The `add-skill` CLI tool has been renamed to `skills`.
Updated all references and command examples in `README.md` and
`CONTRIBUTING.md` to use the new `skills` command.
Specifically:
* `npx add-skill` commands are now `npx skills add`
Rewrote both skill descriptions to eliminate overlap and improve trigger accuracy:
- base44-sdk: Now clearly targets EXISTING projects and SDK code writing
- base44-cli: Now clearly targets project initialization and CLI commands
- Removed conflicting "ALWAYS ACTIVATE on ANY base44 mention" from both
- Added explicit routing to direct users to the correct skill
- Established clear prerequisite relationship (cli before sdk for new projects)
- Added critical warnings and mandatory reading notes for the `base44 create` command to ensure users follow the correct workflow and template selection.
- Updated template selection section in create.md to clarify when to use `backend-only` vs `backend-and-client`.
- Improved examples and added critical steps for using the `backend-only` template with existing frameworks.
- Emphasized the importance of following instructions to avoid project scaffolding issues.
Sync the base44-cli skill documentation with version 0.0.21 of the CLI.
Changes:
- Updated create command: name parameter is now positional argument instead of --name flag
- Added --skills option to create command (controls AI agent skills installation)
- Added --projectId option to link command (allows linking to specific project)
- Updated all examples and usage documentation to reflect new syntax
- Updated CLI_VERSION from v0.0.17 to v0.0.21
Breaking changes:
- create command now uses: npx base44 create my-app -p . (not -n my-app)
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
- Updated SKILL.md to clarify the synchronization process using git-based change detection.
- Added a new section detailing how the skill works, including input validation and change detection steps.
- Improved instructions for validating the source folder and discovering command structures.
- Enhanced the summary of changes presented to the user, including breaking changes and option modifications.
- Updated CLI_VERSION to reflect the new version after processing changes.
* Update SKILL.md to enhance command and option documentation
- Added sections for command-level and option/argument-level changes
- Detailed comparison examples for command options
- Summarized commands added/removed and option changes
- Updated descriptions for clarity and accuracy
- Recommended manual review for specific changes
* Update SKILL.md to improve CLI command discovery process
- Renamed and expanded sections for better clarity and context
- Added a new step to read the CHANGELOG for recent changes
- Adjusted subsequent steps to reflect the new structure
- Enhanced guidance on identifying command changes and documentation updates
- Add comprehensive TypeScript interfaces for User, LoginResponse, and all parameter types
- Add detailed method signatures with return types
- Document all authentication methods including OAuth provider support
- Add error handling section with common error scenarios
- Add environment availability information (frontend/backend)
- Improve examples with complete registration flow including OTP verification
- Add detailed documentation for password reset and change flows
* docs(create): Clarify Base44 project creation command limitations
- Added important notes regarding the command's functionality, emphasizing that it only creates the `base44/` folder with configuration files and does not initialize `package.json`, install npm dependencies, or create project scaffolding.
- Updated documentation to instruct users to set up their npm project separately before or after running the command.
* docs: Update Base44 CLI and SDK installation instructions
- Added installation commands for both Base44 CLI and SDK, emphasizing the importance of not hardcoding package versions to ensure users always get the latest updates.
- Included a note in both SKILL.md files to guide users on proper installation practices.
* fix
- Introduced new documentation for the Sync CLI Skill, detailing the synchronization process with the Base44 CLI source code.
- Updated SKILL.md to include new command references and improved descriptions for existing commands.
- Added new reference files for `dashboard`, `link`, and updated `deploy` and `create` commands to clarify usage and options.
- Enhanced troubleshooting sections and provided examples for better user guidance.
- Add boolean rule types (true/false) for RLS
- Add common RLS patterns section (public create/admin read, owner-only)
- Add complete Field Level Security (FLS) documentation
- Add rls-examples.md with real-world scenarios:
- Todo app, subscription system, social app
- Dating app, enterprise system, mini social network
- Remove deprecated "write" rule mention from RLS (use create/read/update/delete)
- Update SDK entities.md with RLS/FLS overview
- Clarify that asServiceRole does NOT bypass RLS (just sets role=admin)
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
- Clarified that Base44 functions run on Deno and must use `Deno.serve()`.
- Updated examples to reflect new response handling with `Response.json()`.
- Added error handling and input validation in function examples.
- Included information on using service role access and environment variables.
- Updated SKILL.md to include new sections for function management, installation commands, and troubleshooting for functions.
- Added references for creating and deploying functions in new functions-create.md and functions-deploy.md files.
- Improved directory structure examples and clarified function configuration requirements.
github-actions[bot]