feat: implement enhanced bounty discovery algorithm with multi-factor scoring (#3)

[sungdark]

Summary

Implemented an enhanced ranking algorithm for bounty discovery that replaces the basic scoring function with a sophisticated multi-factor scoring system.

Scoring Components

Factor	Max Points	Description
Funding	0-25	Rewards well-funded bounties (80%+ = max), uses diminishing returns curve
Activity	0-20	Rewards active bounties (5 pts per claim, max 4 claims)
Recency	0-20	Exponential decay over 14 days to surface fresh opportunities
Reward	0-20	Logarithmic scale to avoid extreme reward bias
Competition	-15	Penalty based on existing claims to help underrated bounties surface

UI Improvements

• Rank badges (🥇🥈🥉) for top 3 bounties
• Score breakdown showing each factor's contribution
• Better visual hierarchy with improved card layout
• Grid layout (3 columns on large screens)

Example Score Breakdown

For a bounty with 90% funding, 4 claims, 2 days old, $500 reward:

• +Funding: 25 (max score for 80%+ funding)
• +Activity: 20 (4 claims × 5)
• +Recency: 15 (2 days old)
• +Reward: ~16 (log scale)
• -Competition: -15 (4 claims × 3.75)
• Total: ~61/100

Testing

The algorithm correctly ranks:

1. d1 (Auth reliability) - High funding + activity balances lower reward
2. d3 (Notifications) - Medium funding + reward but older
3. d2 (Performance audit) - Fresh but lower funding/reward

Closes #3

Payment

收款地址：eB51DWp1uECrLZRLsE2cnyZUzfRWvzUzaJzkatTpQV9