Add hireEZ 'Consent-First Meeting Ops' RL environment prototype

[SHAFF622]

Motivation

• Provide a hackathon-ready MVP that fits a Customer Success Engineer delivery focus by prioritising safe, execution-oriented automation over speculative emotion-detection features.
• Frame the problem as a sequential decision / RL-style task (ingest → decide → execute/ask) so the agent must trade off speed, consent, and human approval in realistic revenue-ops workflows.

Description

• Add meeting_ops_env.py, a self-contained RL-style environment that defines MeetingState, discrete Action values, step transitions, and reward functions for CRM writes, Slack coordination, and quiet in-call nudges.
• Provide a simple heuristic policy in heuristic_policy and an evaluation harness evaluate_policy(...) that prints summary metrics for demo runs.
• Add README.md explaining the hackathon framing, environment design, demo talking points, and extension ideas like mock adapters and training loops.
• Add unit tests tests/test_meeting_ops_env.py that verify key governance and workflow rules (unsafe CRM writes are penalized, missing fields trigger clarification, and low-confidence cases trigger approval gating).

Testing

• Ran the demo evaluation with python3 meeting_ops_env.py, which prints aggregated metrics (example output shows average reward and action rates).
• Ran the test suite with python3 -m unittest discover -s tests -p 'test_*.py', and all tests passed (Ran 3 tests ... OK).

---

Codex Task