Core team auto-detection from commit patterns

[copyleftdev]

Problem

Every report requires manually specifying core team members:
```python
is_core = author in ['sebmarkbage', 'eps1lon', 'josephsavona', ...] # React
is_core = author in ['tim-smart', 'IMax153', 'mikearnaldi', ...] # Effect-TS
is_core = author in ['charliermarsh', 'konstin', 'zanieb', ...] # uv
```

This is error-prone, requires research per repo, and breaks the "no manual input" promise.

Proposal

Auto-detect core team from commit patterns:

Algorithm:

1. Compute author commit counts
2. Apply Pareto detection: authors above the 80th percentile by commit count = "core"
3. Refine with email domain clustering (same domain = same org)
4. Refine with temporal consistency (active >50% of months = core)
5. Bot detection via `[bot]` suffix or known bot patterns

Output a `core_team` field in governance analysis:
```json
{
"core_team": ["konstin", "zanieb", "charliermarsh"],
"bots": ["dependabot[bot]", "github-actions[bot]"],
"community": ["user1", "user2", ...],
"detection_method": "pareto_80pct + email_domain + temporal_consistency"
}
```

Confidence labels:

• Definite: email domain matches (e.g., all @meta.com = Meta core)
• Dominant: Pareto + temporal consistency
• Heuristic: Pareto only

Evidence

Manually identified core teams for 3 repos. Email domains would have caught all 3.

Complexity

Medium. Extension to governance analysis. The email domain heuristic is powerful.