TeckAgents is an AI agent development repository by WrighTeck, focused on designing, building, and documenting goal-driven AI agents for technology education and workflow automation.
Unlike traditional AI assistants that respond to individual prompts, TeckAgents explores autonomous and semi-autonomous agents that can plan tasks, execute multi-step workflows, use tools, maintain state, and evaluate outcomes.
The goals of TeckAgents are to:
- Learn and apply modern AI agent frameworks
- Design goal-oriented, workflow-based agents
- Explore multi-agent collaboration patterns
- Build agents that educate, guide, and execute
- Document architectural decisions and lessons learned
- Create reusable foundations for future WrighTeck AI products
This repository serves as both a learning lab and a technical portfolio.
Agents developed in this repository will target practical education and execution in the following domains:
- Coding & Software Development
- Software Testing & QA
- AI & Automation
- Tech Productivity
- Troubleshooting & Debugging
- Cybersecurity & Privacy (defensive and educational)
Agents are designed to move beyond explanations and actively guide users through real-world tasks and learning workflows.
Planned and in-progress agent capabilities include:
- Goal decomposition and task planning
- Multi-step execution and iteration
- Tool usage (code, files, APIs, documentation)
- State management and decision branching
- Role-based collaboration (instructor, reviewer, executor)
- Learning path generation and progress evaluation
- Structured outputs (lessons, checklists, test cases, reports)
- Python (primary language for agent logic and orchestration)
- CrewAI – role-based, task-oriented multi-agent systems
- LangGraph – stateful, graph-based agent workflows
- AutoGen – conversational and collaborative agent patterns
- MCP (Model Context Protocol) – standardized tool and context integration
- Async Python (
async/await) - Environment configuration (
.env) - File and data handling
- API integrations
- Logging and observability (as learning progresses)
- Prompt design and agent reasoning strategies
This repository is model-agnostic by design.
Agents may be tested with different LLMs as supported by the frameworks used (e.g., OpenAI-compatible models, open-source models, or future integrations).
The focus is on agent architecture, behavior, and orchestration, not model lock-in.
TeckAgents/
├── README.md
├── notes/
│ └── learning-notes.md
└── experiments/