- Add version numbers, changelogs, and Last Reviewed dates to: - ENVIRONMENT_SETUP.md - SCENARIO_STAGE_MAPPING.md - OPERATIONS_AUDIT.md - training_schema.json - Update all to version 2.0 with changelog entries
6.3 KiB
Hermes Agent Training Learning Prompt
This document is a structured learning prompt for Hermes to learn the AITBC agent training curriculum.
Learning Objective
Hermes should master the complete AITBC agent training program, understanding all 9 stages of agent development from foundation to multi-chain architecture.
Learning Process
Phase 1: Document Analysis
-
Read and analyze all documentation in
/docs/agent-training/:README.md- Overview and structuretraining_schema.json- Schema definitionENVIRONMENT_SETUP.md- Environment configurationOPERATIONS_AUDIT.md- Operations coverageSCENARIO_STAGE_MAPPING.md- Practical applications- All stage JSON files (
stage1_*.jsonthroughstage9_*.json)
-
Identify key patterns across all stages:
- Command structure and parameters
- Validation criteria
- Expected outcomes
- Prerequisites and dependencies
Phase 2: Stage Mastery
For each of the 9 training stages, Hermes should:
- Understand the stage objectives
- Analyze the command sequences
- Identify validation criteria
- Map to practical scenarios
- Note potential failure points
Phase 3: Debugging and Feedback
After analyzing each stage, Hermes should provide:
Debug Messages
- Missing information: What data is incomplete or unclear?
- Inconsistencies: Are there contradictions between stages?
- Ambiguities: What needs clarification?
- Gaps: What's missing from the training curriculum?
Suggestions
- Improvements: How can the training be enhanced?
- Optimizations: What can be streamlined?
- Additions: What topics should be covered?
- Restructuring: How can the curriculum be better organized?
Issue Identification
- Potential failures: What might go wrong during execution?
- Dependency issues: Are there missing prerequisites?
- Environment problems: What setup issues might occur?
- Validation gaps: Are success criteria sufficient?
Stage-by-Stage Learning Guide
Stage 1: Foundation
Focus: Wallet creation, basic transactions, mining, balance verification Key Learning: Core blockchain interaction patterns Debug Focus: CLI command syntax, wallet naming conventions, transaction signing
Stage 2: Operations Mastery
Focus: Advanced wallet operations, transaction monitoring, blockchain queries Key Learning: State management and transaction tracking Debug Focus: Query patterns, monitoring commands, multi-wallet coordination
Stage 3: AI Operations
Focus: Job submission, task management, result retrieval Key Learning: Agent coordination and AI task orchestration Debug Focus: API interactions, task lifecycle management, result parsing
Stage 4: Marketplace Economics
Focus: Marketplace interaction, resource trading, price discovery Key Learning: Economic modeling and market operations Debug Focus: Trading mechanics, price feeds, resource allocation
Stage 5: Expert Operations
Focus: Advanced agent behaviors, autonomous decision making Key Learning: Complex task orchestration and performance optimization Debug Focus: Decision logic, performance metrics, optimization strategies
Stage 6: Agent Identity SDK
Focus: Identity management, authentication, permissions Key Learning: Security protocols and identity workflows Debug Focus: Auth flows, permission checks, security edge cases
Stage 7: Cross-Node Training
Focus: Multi-node coordination, distributed consensus Key Learning: Network topology and failover handling Debug Focus: Node communication, consensus mechanisms, network partitions
Stage 8: Advanced Agent Specialization
Focus: Domain-specific capabilities, expert systems Key Learning: Specialized knowledge bases and custom skills Debug Focus: Knowledge integration, skill development, domain logic
Stage 9: Multi-Chain Architecture
Focus: Cross-chain operations, bridge protocols, interoperability Key Learning: Multi-chain asset management and bridge patterns Debug Focus: Bridge mechanics, cross-chain transactions, asset mapping
Output Format
For each stage, Hermes should provide:
## Stage Analysis: [Stage Name]
### Understanding Summary
[Brief summary of what Hermes learned]
### Debug Messages
- **Issue**: [Description]
- **Impact**: [Why it matters]
- **Suggested Fix**: [How to resolve]
### Suggestions
- **Improvement**: [Description]
- **Rationale**: [Why it helps]
- **Implementation**: [How to implement]
### Potential Failures
- **Failure Scenario**: [Description]
- **Detection**: [How to identify]
- **Mitigation**: [How to prevent]
Final Deliverable
After completing all 9 stages, Hermes should provide:
- Comprehensive Learning Summary: What was learned across all stages
- Cross-Stage Analysis: Patterns and relationships between stages
- Curriculum Assessment: Overall quality and completeness
- Implementation Roadmap: How to apply this knowledge in practice
- Testing Strategy: How to validate agent training success
Interactive Feedback Protocol
When Hermes encounters issues during learning:
- Stop and document the specific issue
- Provide context from the relevant documentation
- Suggest alternatives or workarounds
- Request clarification if information is missing
- Propose improvements to the documentation
Success Criteria
Hermes has successfully learned when it can:
- Explain each stage's objectives and methods
- Identify potential issues before they occur
- Suggest improvements to the training curriculum
- Debug training execution problems
- Provide actionable feedback for curriculum enhancement
Instructions for Use:
- Submit this prompt to Hermes along with access to
/docs/agent-training/ - Request Hermes to follow the learning process systematically
- Collect Hermes's stage-by-stage analysis and feedback
- Use the feedback to improve the training curriculum
- Iterate based on Hermes's suggestions
Last Updated: 2026-05-09 Version: 2.0 Purpose: Hermes AI agent learning prompt for AITBC agent training curriculum Changelog:
- v2.0 (2026-05-09): Added concrete examples, templates, and detail level guidelines
- v1.0 (2026-05-07): Initial version with basic learning process structure