[Feature]: Create `support_ticket_agent` Example Template

[rishuraj1]

Problem Statement

The project documentation (README, Environment Setup, and Developer Guide) consistently references a support_ticket_agent as the primary example for verifying installation. However, this agent is not bundled with the repository, and the exports/ directory where it is expected to exist is gitignored and empty. This creates a mismatch between the documented onboarding flow and the actual user experience.

Proposed Solution

Create a minimal example agent at:
exports/support_ticket_agent/

Alternatives Considered

Documenting that no default agent is provided
This shifts the burden onto users and does not fix the disconnect with existing docs.

Linking to an external support_ticket_agent template repo
Adds an extra step during onboarding and breaks offline setup.

Generating the agent dynamically via a CLI command
Could be a future enhancement but introduces additional complexity for beginners.

Additional Context

The README currently states that users can run a validation command against the support_ticket_agent, implying it already exists.

ENVIRONMENT_SETUP.md suggests using this agent to verify LLM connectivity in both mock and real modes.

The exports/ directory is present but empty due to .gitignore, which gives the impression that something is missing rather than intentionally excluded.

Having a concrete example aligns with common patterns in AI agent frameworks (e.g., LangChain, Semantic Kernel, AutoGen), all of which include working starter agents.

This would serve as the canonical reference for how to lay out an agent, which is not obvious given the current folder structure and tooling.

Implementation Ideas

Create a cookie-cutter minimal agent structure and commit it to exports/support_ticket_agent/.

Default the agent’s agent.json to a three-step linear graph with mock-friendly node descriptions.

Keep tools.py self-contained with no external dependencies to maintain simplicity.

Leverage AgentRunner.validate() in this agent’s main.py so users can run:

python -m support_ticket_agent validate --mock

Include a local-only README demonstrating:

Running in mock mode

Running with actual LLM providers (optional)

Expected I/O examples

Update existing docs to reference this agent instead of a hypothetical one.

Consider extending the CLI later to auto-generate new agents based on this template.

[rishuraj1]

I would like to work on this issue by implementing the above minimal support_ticket_agent template, wiring it to work correctly in both mock and real LLM modes, and updating the documentation where necessary to ensure the onboarding flow reflects the availability of this example agent.

[github-actions]

Found a possible duplicate of #696: Both issues address the missing exports/ directory that's referenced in documentation. Issue #696 covers the broader problem (missing exports/ directory structure), while this issue proposes a specific solution (creating a support_ticket_agent example). Consider tracking this as a sub-task of #696 or consolidating efforts there.

If you disagree with the duplicate assessment, please react with 👎