emerging-movers
npx skills add https://github.com/senpi-ai/senpi-skills --skill emerging-movers
Agent 安装分布
Skill 文档
Emerging Movers Detector v3.1
Tracks Smart Money market concentration across all Hyperliquid assets and flags assets accelerating up the ranks before they become crowded top-3 plays. By the time an asset hits the top of the SM leaderboard, the easy money is gone. This catches the trajectory.
One API call per scan. Near-zero LLM tokens. Runs every 60 seconds.
How It Works
The SM Profit Concentration Leaderboard
Senpi’s leaderboard_get_markets returns all assets ranked by percentage of total Smart Money profit in the last 4-hour rolling window. This isn’t trader count â it’s where the money is actually flowing.
#1 ETH SHORT 31.4% 286 traders
#2 BTC SHORT 25.1% 436 traders
#3 HYPE SHORT 24.2% 330 traders
...
#36 ASTER SHORT 0.2% 18 traders â 60s later: #13, 0.82%, 65 traders
The script tracks this leaderboard over time and detects acceleration.
Detection Signals
Immediate Action Signals (v3+)
| Signal | Condition | Priority |
|---|---|---|
| IMMEDIATE_MOVER | 10+ rank jump from #25+ in ONE scan | Highest â act now |
| NEW_ENTRY_DEEP | Appears in top 20 from nowhere | Very high |
| CONTRIB_EXPLOSION | 3x+ contribution increase in one scan | Very high |
| DEEP_CLIMBER | 5+ rank jump from #25+ | High |
Trend Signals
| Signal | Condition |
|---|---|
| NEW_ENTRY | First appearance in top 50 |
| RANK_UP | Jumped 2+ positions in one scan |
| CLIMBING | 3+ positions up over several scans |
| ACCEL | Contribution % increasing scan-over-scan |
| STREAK | Consistently climbing every check |
| VELOCITY | Sustained positive contribution growth |
v3.1 Quality Filters
These prevent false IMMEDIATE signals that looked great on rank jump alone but failed on execution:
| Filter | Rule | Rationale |
|---|---|---|
| Erratic rank | >5 rank reversals in history â erratic: true, downgraded |
Bouncing ranks are noise |
| Velocity gate | contribVelocity < 0.03 â lowVelocity: true, excluded from IMMEDIATE |
No momentum behind the move |
| Trader count floor | <10 traders â SKIP IMMEDIATE | Single whale risk |
| Max leverage check | max leverage < 10x â SKIP | Not worth the limited position sizing |
See references/quality-filters.md for implementation details and real-world examples.
Architecture
ââââââââââââââââââââââââââââââââââââââ
â Cron: every 60 seconds â
ââââââââââââââââââââââââââââââââââââââ¤
â scripts/emerging-movers.py â
â ⢠Loads scan history from JSON â
â ⢠Fetches leaderboard (1 API call) â
â ⢠Parses top 50 markets â
â ⢠Compares with previous scans â
â ⢠Detects signals + v3.1 filters â
â ⢠Saves updated history â
â ⢠Outputs JSON with alerts â
ââââââââââââââââââââââââââââââââââââââ¤
â Agent reads output: â
â ⢠IMMEDIATE alerts â evaluate now â
â ⢠Deep climbers â queue for review â
â ⢠No alerts â silent â
ââââââââââââââââââââââââââââââââââââââ
Files
| File | Purpose |
|---|---|
scripts/emerging-movers.py |
Scanner script |
emerging-movers-history.json |
Auto-managed scan history (last 60 scans) |
max-leverage.json |
Optional: asset max leverage reference |
Output
See references/output-schema.md for the complete JSON schema.
Key top-level fields: alerts[], topMovers[], immediateMovers[], deepClimbers[], scanCount, timestamp.
Per-alert fields: asset, direction, rank, prevRank, contribution, traderCount, reasons[], contribVelocity, isImmediate, isDeepClimber, erratic, lowVelocity.
Cron Setup
*/1 * * * * python3 scripts/emerging-movers.py
Agent Response Logic
isImmediate: true+erratic: false+lowVelocity: falseâ Evaluate immediately for entry via ScannerisDeepClimber: trueâ Queue for next scanner runerratic: trueorlowVelocity: trueâ Log but do not act- No alerts â Silent
Companion Recipes
- opportunity-scanner â use Scanner to deep-dive assets flagged by Emerging Movers
- autonomous-trading â full loop integrating Emerging Movers as entry trigger
- wolf-strategy â uses IMMEDIATE_MOVER as primary entry signal