Issue Link (REQUIRED)

Fixes #119

Summary

Changes

Adds python/src/agent_squad/agent_overlap_analyzer.py — a pure-stdlib Python implementation of the existing TypeScript AgentOverlapAnalyzer.

Algorithm (matches the TypeScript reference):

1. Tokenise and remove English stopwords from each agent description
2. Build TF-IDF vectors (implemented without external dependencies)
3. Compute pairwise cosine similarity
4. Label pairs: High (>0.3), Medium (>0.1), Low (≤0.1)
5. Derive a uniqueness score per agent (1 − mean similarity to others)

Public API mirrors the TypeScript class:

from agent_squad.agent_overlap_analyzer import AgentOverlapAnalyzer

analyzer = AgentOverlapAnalyzer({
    "flight_agent": {"name": "Flight Agent", "description": "..."},
    "hotel_agent":  {"name": "Hotel Agent",  "description": "..."},
})
result = analyzer.analyze_overlap()
# result.pairwise_overlap["flight_agent__hotel_agent"].potential_conflict
# → "Low" | "Medium" | "High"

Also adds python/src/tests/test_agent_overlap_analyzer.py (12 tests) covering: return type, pair count, key format, percentage bounds, conflict levels, uniqueness score count and bounds, high/low overlap cases, single-agent and empty-agent edge cases.

No new external dependencies — pure stdlib (math, re).

User experience

Before: Python users had no equivalent of the TypeScript overlap analyzer and had to write their own or use the TS version.

After: same analysis is available in Python with identical thresholds and output format.

Checklist

• I have performed a self-review of this change
• Changes have been tested
• Changes are documented
• I have linked this PR to an existing issue (required)

Acknowledgment

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.