- Add currency field (AIT) to all pricing parameters (market_gpu_register, market_bid, market_sell)
- Fix market_bid to reference gpu_id instead of listing_id (GPU-specific)
- Fix market_buy to reference gpu_id instead of listing_id for consistency
- Add market_my-listings operation to check seller's own listings
- Add economics_model operation for price prediction (model_type, features)
- Add wallet_balance check at beginning of marketplace operations
- Add stage4_marketplace_economics to depends_on (AI operations need GPU listings from marketplace)
- Make ai_training_agent operation optional (gpu_listing may not exist)
- Add resource_depends field to training_schema.json for cross-stage resource dependencies
- Add ai_jobs check after ai_submit to verify job appears in listing
- Add finetune_status polling after finetune_submit for job monitoring
- Add depends_on validation documentation to training_schema.json (verify prerequisite stages mastered)
- Add performance validation to network_ping (max_latency_ms: 100)
- Add wallet_balance check after wallet_send to verify transaction deduction
- Add mining_rewards verification after mining_stop to confirm rewards collected
- Move wallet_delete to end of operations array to prevent subsequent operations from failing
- Add curriculum fields to stages 7, 8, 9 (difficulty, skill_level, depends_on, skills, objectives)
- Add missing agent types to README (specialized, architect)
- Fix schema inconsistency in ENVIRONMENT_SETUP.md - update example to use official training_schema.json format
- Fix Stage 1 integration test - remove wallet_fund from operations, use existing operations
- Make messaging_send operation optional in Stage 1 (messaging configuration is optional per ENVIRONMENT_SETUP.md)
- Update scenario references in all stage files to point to correct directory (/docs/scenarios/)
- Add agent type definitions section to README.md (coordinator, genesis, follower, wallet, general)
- Clarify setup_method in Training Schema section - stages use Python-based setup system (aitbc.training_setup module) rather than individual shell scripts
- Fix Python version requirement from 3.10+ to 3.13.5 in Prerequisites
- Add skills field to training schema
- Add skill_level field to training schema (novice/apprentice/practitioner/expert/master)
- Add objectives field to training schema
- Add certifications field to training schema
- Add curriculum fields to stage1_foundation.json (skills, skill_level, objectives, certification)
- Add curriculum fields to stage2_operations_mastery.json (skills, skill_level, objectives, certification)
- Add curriculum fields to stage3_ai_operations.json (skills, skill_level, objectives)
- Add curriculum fields to stage4_marketplace_economics.json (skills, skill_level, objectives)
- Add curriculum fields to stage5_expert_operations.json (skills, skill_level, objectives)
- Add curriculum fields to stage6_agent_identity_sdk.json (skills, skill_level, objectives)
- Add integration_tests field to validation section
- Add coverage_target field for test coverage percentage
- Add coverage_report field for coverage reporting
- Add mock_data section with wallets, addresses, transactions
- Add integration test example to stage1_foundation.json
- Add coverage target and report to stage1_foundation.json
- Add mock data example to stage1_foundation.json
- Add examples field to operation schema (CLI command examples)
- Add errors field to operation schema (expected errors and resolution)
- Add verify field to operation schema (verification commands)
- Add resources field to operation schema (learning resource links)
- Add documentation examples to wallet_create operation in stage1_foundation.json
- Add idempotent field to operation schema
- Add check_exists field for resource existence validation
- Add verify_on_chain flag for transaction verification
- Add confirmations_required field for block confirmations
- Add timeout_seconds field for on-chain verification timeout
- Add retry configuration (max_attempts, backoff_seconds, retryable_errors)
- Add max_memory_mb to performance metrics
- Add validation examples to wallet_create operation (idempotent, check_exists, performance)
- Add validation examples to transaction_send operation (verify_on_chain, retry, performance)
- Add depends_on field to training schema for stage dependencies
- Add difficulty field to training schema (beginner/intermediate/advanced/expert)
- Add estimated_time_minutes field to training schema
- Add structural fields to stage1_foundation.json (beginner, 30min, no dependencies)
- Add structural fields to stage2_operations_mastery.json (beginner, 45min, depends on stage1)
- Add structural fields to stage3_ai_operations.json (intermediate, 60min, depends on stage2)
- Add structural fields to stage4_marketplace_economics.json (intermediate, 60min, depends on stage2)
- Add structural fields to stage5_expert_operations.json (advanced, 90min, depends on stage3, stage4)
- Add structural fields to stage6_agent_identity_sdk.json (advanced, 75min, depends on stage5)
- Create HERMES_LEARNING_PROMPT.md with comprehensive learning instructions
- Update README.md with Hermes-specific learning instructions
- Add interactive prompts for Hermes to provide debug messages and suggestions
- Structure learning process across 9 training stages with feedback protocol
- Add aitbc2 blockchain node access documentation via gitea-runner SSH
- Fix wallet send command syntax to use proper --wallet-name flag format
- Update stage1 training commands to include --fee and --password parameters
- Switch from faucet wallet to proposer wallet for funding
- Proposer is the current PoA proposer and valid signer
- Send 1M AIT from genesis to proposer first
- Send 100 AIT from proposer to training-w1
- Increased sleep times to 15s for transaction confirmation
Replaced direct AITBC CLI command execution with OpenClaw agent-based execution that respects the allowlist:
- Changed openclaw_training_operations to execute commands via `openclaw agent --message` instead of direct CLI calls
- Removed operation-specific command building logic (wallet_create, genesis_init, etc.)
- Simplified execution flow to single OpenClaw agent invocation with prompt message
- Added prerequisites
Added comprehensive agent training functionality that executes actual AITBC CLI commands:
- Renamed openclaw_operations to openclaw_training_operations in aitbc_cli.py
- Added train action with agent/validate/certify subcommands to openclaw_operations
- Implemented agent training that loads JSON training data and executes real CLI commands
- Added operation mapping for wallet, blockchain, messaging, and system commands
- Skip
