- Add conn.commit() to agent registration in agent-registry
- Remove unused integration_layer.py and coordinator.py from agent-services
- Fix blockchain RPC endpoint from /rpc/sync to /rpc/syncStatus
- Replace Annotated[Session, Depends(get_session)] with Session = Depends(get_session) for cleaner dependency injection syntax across marketplace routers
- Add transaction data to RPC responses (get_block, get_blocks_range)
- Fix import_block to handle transactions parameter
- Change database paths to absolute paths for consistency
- Make dev_heartbeat.py executable
🔐 Guardian Contract Security Enhancements:
- Add persistent SQLite storage for spending history and pending operations
- Replace in-memory state with database-backed state management
- Implement _init_storage(), _load_state(), _save_state() for state persistence
- Add _load_spending_history(), _save_spending_record() for transaction tracking
- Add _load_pending_operations(), _save_pending_operation(), _remove_pending_operation()
- Add production genesis initialization scripts
- Add keystore management for production
- Add production node runner
- Add setup production automation
- Add AI memory system for development tracking
- Add translation cache service
- Add development heartbeat monitoring
- Update blockchain RPC router
- Update coordinator API main configuration
- Update secure pickle service
- Update claim task script
- Update blockchain service configuration
- Update gitignore for production files
Resolves conflicts by accepting PR branch changes
- Add production genesis initialization scripts
- Add keystore management for production
- Add production node runner
- Add setup production automation
- Add AI memory system for development tracking
- Add translation cache service
- Add development heartbeat monitoring
- Update blockchain RPC router
- Update coordinator API main configuration
- Update secure pickle service
- Update claim task script
- Update blockchain service configuration
- Update gitignore for production files
- Introduced RestrictedUnpickler in secure_pickle.py to prevent arbitrary code execution
- Updated IPFSStorageService.retrieve_memory and decompress_memory to use safe_loads()
- Maintains pickle dumps for serialization (safe)
- Reduces risk of remote code execution via malicious pickled data
- Create ai.py with serve and request commands
- request includes balance verification and payment via blockchain send
- serve runs FastAPI server and optionally registers jobs with coordinator
Update marketplace.py:
- Add gpu unregister command (DELETE endpoint)
- Drop tables before create in development
- Catch OperationalError for duplicate indexes
- Add logging for errors
This allows the Coordinator API to start cleanly with SQLite even if previous runs left residual schema.
Also adds debugging status document for aitbc1 branch.
Fixes startup failure on fresh deployment.
- Add multi-chain support with ait-devnet, ait-testnet, ait-mainnet
- Add chain selector dropdown in explorer UI
- Add /api/chains endpoint to list supported chains
- Update all API endpoints to accept optional chain_id parameter
- Update RPC URL configuration to support multiple chains per network
- Change default port from 3001 to 8016 (enhanced services range)
- Update coordinator database path to ./
- Add GET /state endpoint to blockchain RPC router for chain state information
- Add GET /rpc/getBalance/{address} endpoint for account balance queries
- Add GET /rpc/head endpoint to retrieve current chain head block
- Add GET /rpc/transactions endpoint for latest transaction listing
- Add chain-specific wallet balance endpoint to wallet daemon
- Add blockchain state CLI command with --all-chains flag for multi-chain queries
- Add CLI interface functions to enterprise_integration.py: create_tenant, get_tenant_info, generate_api_key, register_integration, get_system_status, list_tenants, list_integrations
- Replace direct service imports with importlib-based module loading to avoid naming conflicts
- Refactor start_gateway command to create_tenant_cmd with name and domain parameters
- Update integration test success rate from
- Add multi-chain configuration with devnet, testnet, and mainnet RPC URLs
- Add chain selector dropdown in explorer UI for network switching
- Add chain_id parameter to all API endpoints (chain/head, blocks, transactions, search)
- Add /api/chains endpoint to list supported blockchain networks
- Update blockchain explorer port from 3001 to 8016
- Update devnet RPC port from 8080 to 8026
- Add GPU reviews table
- Replace SessionDep type alias with explicit Annotated[Session, Depends(get_session)]
- Add missing imports for Session, Annotated, and Depends types
- Update all endpoint function signatures to use explicit dependency annotation
- Remove redundant `= Depends()` default values from session parameters
- Update docstrings and comments to reference new annotation pattern
- Apply changes consistently across all router
- Replace SessionDep type alias with explicit Annotated[Session, Depends(get_session)]
- Add missing imports for Session and Annotated types
- Update all endpoint function signatures to use explicit dependency annotation
- Apply changes consistently across all router files (admin, agent, marketplace, etc.)
- Add marketplace_gpu router to main.py router includes
- Add POST /marketplace/gpu/{gpu_id}/confirm endpoint for client booking acknowledgment
- Add POST /tasks/ollama endpoint for Ollama task submission with model and prompt parameters
- Add POST /payments/send endpoint as stub for blockchain payment processing
- Add GPUConfirmRequest, OllamaTaskRequest, and PaymentRequest models
- Add CLI commands: confirm, ollama-task, and pay for new endpoints
- Include validation
- Add validation for negative and zero booking hours
- Add maximum booking duration limit (8760 hours = 1 year)
- Add validation to ensure booking end time is in future
- Prevent negative costs and invalid booking periods
- Improve error messages with detailed validation feedback
Fixes edge cases where users could book GPUs with invalid parameters
leading to negative costs and impossible booking periods.
- Change database from in-memory to file-based SQLite at aitbc_coordinator.db
- Add status="active" to GPU booking creation
- Allow GPU release even when not properly booked (cleanup case)
- Add error handling for missing booking attributes during refund calculation
- Fix get_gpu_reviews query to use scalars() for proper result handling
- Rename metadata to user_metadata in TenantUser model
- Rename metadata to usage_metadata in UsageRecord model
- Rename metadata to invoice_metadata in Invoice model
- Rename metadata to event_metadata in TenantAuditLog model
- Update TenantManagementService to use event_metadata parameter name
- Change file mode from 644 to 755 for all project files
- Add chain_id parameter to get_balance RPC endpoint with default "ait-devnet"
- Rename Miner.extra_meta_data to extra_metadata for consistency
- Add role detection to command groups (admin, client, miner, blockchain)
- Load role-specific config files (~/.aitbc/{role}-config.yaml)
- Add role field to Config class with environment variable support
- Implement automatic role detection from invoked subcommand
- Add development mode API key bypass for testing (APP_ENV=dev)
- Update CLI checklist with role-based configuration documentation
- Add configuration override priority and
- Add debug print statements to _validate_api_key and require_admin_key for troubleshooting
- Add /admin/debug-settings and /admin/test-key endpoints for API key validation testing
- Bypass require_admin_key dependency in /admin/stats endpoint for direct validation
- Fix database warmup to properly handle session generator lifecycle
- Re-enable all previously disabled routers in main.py
- Add custom OpenAPI security scheme
- Comment out most routers in main.py to isolate Pydantic issue
- Keep only blockchain router enabled for testing
- Fix database warmup to use get_session() instead of SessionDep()
- Add blockchain router to __init__.py exports
- Add test endpoint to agent_router for verification
- Duplicate agent network and execution receipt endpoints in client and exchange routers as temporary workaround
- Add /v1/admin/status endpoint for system status
- Add /v1/blockchain/status endpoint for blockchain status
- Add /v1/blockchain/sync-status endpoint for sync status
- Add /v1/monitor/dashboard endpoint for monitoring dashboard
- Fix router imports and missing dependencies
- Handle optional dependencies gracefully (torch, tenseal)
- Update admin router with comprehensive system status endpoint
- Fix blockchain router endpoint paths
- Improve error handling in monitoring dashboard
These endpoints resolve CLI 404/405 errors reported in testing.
- Add get_miner_id() dependency to extract miner ID from X-Miner-ID header
- Update miner register and heartbeat endpoints to require both X-Miner-ID and X-Api-Key headers
- Remove miner_id from query parameters in favor of header-based extraction
- Fix miner heartbeat CLI to send proper JSON payload with inflight, status, and metadata fields
- Fix typo in MinerService: extra_metadata → extra_meta_data
- Add POST /api/v1/miners/{miner_id}/jobs endpoint for listing miner-assigned jobs with filtering
- Add POST /api/v1/miners/{miner_id}/earnings endpoint for miner earnings tracking (mock implementation)
- Add PUT /api/v1/miners/{miner_id}/capabilities endpoint for updating miner capabilities
- Add DELETE /api/v1/miners/{miner_id} endpoint for miner deregistration (sets OFFLINE status)
- Add JobService.fail_job()
- Add /rpc/info endpoint to blockchain node for comprehensive chain information
- Add /rpc/supply endpoint for token supply metrics with genesis parameters
- Add /rpc/validators endpoint to list PoA validators and consensus info
- Add /api/v1/agents/networks endpoint for creating collaborative agent networks
- Add /api/v1/agents/executions/{id}/receipt endpoint for verifiable execution receipts
- Add /api/v1/jobs and /api/v1/jobs/
- Update requirements.txt to make concrete-ml optional with clear comment
- Enhance FHE service warning message to include Python version info
- Remove concrete-ml from mandatory dependencies due to Python 3.13 incompatibility
- Maintain full functionality with TenSEAL provider as default
This resolves the 'Concrete ML not installed' warning by providing clear
information about the version compatibility issue while preserving all
FHE functionality through the TenSEAL provider.
- Update CORS allow_origins in blockchain-node app.py and gossip relay.py to use new port logic (8000-8016)
- Update coordinator-api config.py and config_pg.py with standardized port ranges and service labels
- Update coordinator-api health check script from port 18000 to 8000
- Update geo_load_balancer.py to use configurable host/port (default 0.0.0.0:8017)
- Update agent_security.py sandbox
- Convert _propose_block, _ensure_genesis_block to async methods in PoAProposer
- Add await to gossip_broker.publish calls in block broadcasting
- Add await to _ensure_genesis_block and _propose_block method calls
- Update coordinator-api Python version constraint to >=3.13,<3.15
- Add torch ^2.10.0 dependency to coordinator-api
- Add make_asgi_app import to prometheus_client in main.py
- Export new routers: global
- Update Q3 2026 from "CURRENT PHASE" to "COMPLETED PHASE" with all weeks 13-24 marked complete
- Mark Q4 2026 as "NEXT PHASE" with weeks 25-28 Global Expansion APIs as 🔄 NEXT
- Update priority focus areas from "Next Phase" to "Current Phase" with global expansion emphasis
- Mark Enterprise Integration APIs and Scalability Optimization as ✅ COMPLETE
- Update Phase 4-6 success metrics to ✅ ACHIEVED
- Add PyTorch neural network implementations for PPO, SAC, and Rainbow DQN agents with GPU acceleration
- Implement PPOAgent with actor-critic architecture, clip ratio, and entropy regularization
- Implement SACAgent with separate actor and dual Q-function networks for continuous action spaces
- Implement RainbowDQNAgent with dueling architecture and distributional RL (51 atoms
- Add developer_platform and governance_enhanced router imports to coordinator-api main.py
- Include developer_platform and governance_enhanced routers with /v1 prefix
- Refactor governance_service.py delegation logic to use delegatee_id directly and add logging
- Remove redundant session.add() calls, rely on session.refresh() after commit
- Update governance service docstring to reflect multi
