- Creates database_name column in bots table
- Creates tenant_id column for multi-tenant support
- Adds indexes for both columns
- Migration runs automatically on server start
- Added database_name field to bots schema
- Bot creation now creates a dedicated PostgreSQL database (bot_{name})
- Updated add_bot.rs to create database and store database_name
- Added create_bot_database() function with safe name validation
- Added dynamic table check to all db_api handlers
- Added clean_file_content() to strip markdown fences from LLM responses
- For HTML: extracts content from <!DOCTYPE> to </html>
- For CSS/JS: filters out explanation lines (Here, This, Note:, etc)
- Fixes malformed HTML output from LLM
- Added serve_vendor_file() to serve from {bot}.gblib/vendor/ in MinIO
- Added /js/vendor/* route to app_server
- Removed local ServeDir for /js/vendor from main.rs
- Added ensure_vendor_files_in_minio() to upload htmx.min.js on startup
- Uses include_bytes! to embed htmx.min.js in binary
- Added table_exists_in_database() to check if table exists in PostgreSQL
- Updated validate_table_name() to allow valid identifiers (not just whitelist)
- Added validate_table_name_with_conn() for full validation with DB check
- Added is_table_allowed_with_conn() for handlers to verify table existence
- Updated list_records_handler and count_records_handler to use dynamic check
- Uses parameterized query for table existence check (SQL injection safe)
- Added rewrite_cdn_urls() to replace HTMX CDN URLs with /js/vendor/htmx.min.js
- Applied to both MinIO and filesystem serving paths
- Supports unpkg.com, jsdelivr, cdnjs variants
- Ensures old apps with CDN references work with local files
- Added /js/vendor route to serve local vendor JS files
- Downloaded htmx.min.js v1.9.10 to botserver-stack/static/js/vendor/
- Reverted CSP to strict 'self' only (no external CDN)
- Updated APP_GENERATOR_PROMPT to use /js/vendor/htmx.min.js
- Updated designer prompt to use local HTMX path
- Designer now uses state.bucket_name (like app_generator) instead of DB lookup
- Fixed local file path to match app_server fallback: {site_path}/{bot}.gbai/{bot}.gbapp/{app}/{file}
- Fixed S3 path to match app_server: {bot}.gbapp/{app}/{file} in bucket {bot}.gbai
- Added S3 bucket creation retry logic (like app_generator)
- Updated CSP to allow unpkg.com, cdnjs.cloudflare.com, cdn.jsdelivr.net for scripts/styles
- Added fonts.googleapis.com and fonts.gstatic.com for web fonts
- Updated APP_GENERATOR_PROMPT to use HTMX CDN instead of non-existent /js/vendor path
- Added designer prompt guidelines for relative asset paths
- Remove checkbox text [x]/[>]/[ ] from HTML, use CSS dots only
- Add View Details links to sections
- Add status-indicator (pulsing dot) and status-gear icon
- Auto-expand running sections/children
- Format runtime and estimated time properly
- Fix decision point display
- Fix MIME type for app files by preserving directory structure in sanitize_file_path()
- Add with_event_type() to TaskProgressEvent for correct WebSocket event types
- broadcast_manifest_update() now sends 'manifest_update' type correctly
- update_item_status() broadcasts automatically for real-time file progress
- Update step_results in DB with real file list during generation
- Show app URL in completion event and notification
- Update task progress/current_step/total_steps as files are written
- Mark task as completed with app_url when done
- Fix MinIO health check to use HTTPS instead of HTTP
- Add Vault connectivity check before fetching credentials
- Add CA cert configuration for S3 client
- Add Qdrant vector_db setup with TLS configuration
- Fix Qdrant default URL to use HTTPS
- Always sync templates to S3 buckets (not just on create)
- Skip .gbkb root files, only index files in subfolders
- Merge all 6.0.x through 6.1.x migrations into one file
- All 129 CREATE TABLE statements preserved
- 3661 lines of SQL in consolidated up.sql
- No tables removed, just combined into single migration
Database Schema v7.0.0:
- Create new 'gb' schema with PostgreSQL ENUMs instead of VARCHAR for all domain values
- Add sharding infrastructure (shard_config, tenant_shard_map tables)
- Implement partitioned tables for sessions, messages, and analytics (monthly partitions)
- Add Snowflake-like ID generation for distributed systems
- Design for billion-user scale with proper indexing strategies
Rust Enums:
- Add comprehensive enum types in core/shared/enums.rs
- Implement ToSql/FromSql for Diesel ORM integration
- Include: ChannelType, MessageRole, MessageType, LlmProvider, ContextProvider
- Include: TaskStatus, TaskPriority, ExecutionMode, RiskLevel, ApprovalStatus, IntentType
- All enums stored as SMALLINT for efficiency
Other fixes:
- Fix hardcoded gpt-4 model in auto_task modules to use bot config
- Add vector_db to required bootstrap components
- Add Qdrant health check before KB indexing
- Change verbose START messages to trace level
- Fix episodic memory role handling in Claude client
- Disable auth for /api routes during development
This is a DESTRUCTIVE migration - only for fresh installations.
- Change from error to warn when bucket creation fails
- Continue bootstrap without drive if MinIO not available
- Prevents startup failure when S3 not configured
- Move app_generator, intent_classifier, intent_compiler, autotask_api, designer_ai, ask_later, auto_task, safety_layer to src/auto_task/
- Create auto_task/mod.rs with exports and route configuration
- Update imports in moved files
- Update main.rs to use auto_task::configure_autotask_routes
- Keep table_definition in keywords (shared utility)
- Add execute_task_handler for POST /api/autotask/:task_id/execute
- Add get_task_logs_handler for GET /api/autotask/:task_id/logs
- Add apply_recommendation_handler for POST /api/autotask/recommendations/:rec_id/apply
- Update route configuration in mod.rs
