Agent Skills.
Plug & Play.
Get Started
Setup Guide
Choose your integration method and start building autonomous trading strategies in minutes.
Install the Skill
Install from ClawHub or ask your agent to load the skill file directly.
Try These Prompts
Start trading by giving your agent natural language instructions.
Beyond Simple Strategies
Deep Use Cases
Leveraging agent intelligence, applied trading system, more advanced use cases. Real capital, real results — on Hyperliquid.
Self-Improving Contrarian
CONTRARIAN · 2h before market open & 2h after close (Weekends included)Loading wallet data…
> I will provide you with two wallets, each with 200 USDC starting fund, build a CONTRARIAN personality. You trade twice a day, 2 hours before market opens and 2 hours after market close (include weekends). Use Superior Trade tools to build profitable strategies. Use all your available skills and API access to achieve a yearly returns that is better than S&P500.
The best practice to avoid failing strategy is to always backtest before deploying, it is likely that a backtest with trades in a likely trading time range is a working strategy.
You will only interact with two markdown files - TRADING_JOURNAL.md and ERROR.md
TRADING_JOURNAL.md:
- Your persona and settings
- Current strategies and reasoning
- Trading history & lesson learnt, with statistics like P&L
ERROR.md:
- If you face any technical errors, append a detailed logs, from newest to latest
Since this is the first time we set up, initialize everything and start a strategy immediately for the first time.Capabilities
From Experiment
to Execution
Everything we've learned — every hard-won insight about strategy formulation, code execution, backtesting, and live deployment — available as an agentic platform.
Intent → Execution, end-to-end
Turn an objective into a strategy your agent can run—defined rules, defined risk, versioned end-to-end.
- Intent → strategy blueprint (signals, sizing, guardrails)
- Templates + parameters for fast iteration
- Promote safely: draft → simulate → go live
Simulate first, deploy second
Stress-test scenarios before touching capital—previews, constraints, and mode parity.
- Read-only previews for fills, slippage, and limits
- Scenario sweeps (fees, sizing, volatility, leverage)
- Paper/live parity checks so simulations match deployment behavior
Production-grade live deployment
Deploy strategies like services: controlled lifecycle, secure execution, and runtime monitoring.
- Wallet/credential handling with scoped access
- Start/stop/pause/resume with safe state management
- Live logs + health indicators (errors, latency, drift)
Execution quality you can measure
See what execution actually costs and how clean the fills are—per trade, per strategy.
- Order controls: limit/market, post-only, reduce-only
- Slippage tracking + execution reports
- Cost breakdown: fees, funding, impact (where measurable)
Risk guardrails & circuit breakers
Keep strategies inside boundaries—caps, stops, and auto-pauses when conditions break.
- Strategy SL/TP envelopes + max-loss rules
- Exposure caps (per asset and portfolio)
- Auto-pause triggers for spread/slippage/margin stress
Full decision traceability
Know exactly what the agent did and why—replayable timelines with receipts.
- Action logs: reasoning + latency + outcomes
- Strategy timelines (config changes, trades, incidents)
- Exportable audit packs for debugging and review
Trader graph intelligence
Find traders worth following with real filters—returns, risk, and consistency.
- Rank by ROI, PnL, win rate, drawdown, gain-to-pain
- Behavior tags: activity tempo, consistency, risk posture
- Drift detection when a trader’s regime changes
“Who’s hot” flow & momentum radar
Spot what’s working now and what’s driving it—fast leaderboards with attribution.
- Rolling momentum boards (choose your window)
- Tiered event triggers by notional significance
- Market attribution + concentration scoring
Mirror trading as a managed service
Copy traders with controls—scaling, slippage rules, protections, and clean exits.
- Create/fund/configure mirror strategies in one flow
- Multipliers + slippage policy + strategy-level protection
- Selective per-asset unwinds without killing the strategy
Market data fabric (fast + complete)
All the market structure data your agent needs to generate signals and monitor execution.
- Candles + order book depth
- Funding history, open interest, volume
- Batch scans across the full perp universe
Portfolio & treasury operations
Understand exposure instantly and move capital cleanly when needed.
- Balances, withdrawables, live positions
- Allocation map + PnL summaries
- USDC bridge-out workflow to supported networks
Agent-native automation hooks
React to events automatically—alerts, triggers, and scheduled loops.
- Webhooks/events for thresholds and state changes
- Alert routing (Discord/Slack/bot)
- Schedulers for scans, rebalances, recurring tasks
Strategy versioning & controlled rollout
Ship changes safely—version, stage, roll forward or roll back.
- Versioned strategies with change notes
- Hot edits with rollback support
- Sandbox staging before promoting live
Docs, playbooks & embedded operator help
Keep guidance inside the platform so agents and operators don’t context-switch.
- Built-in guide catalog + searchable reference
- Best-practice snippets + runbooks
- Quickstarts for common agent workflows
Watchlists that behave like a live feed
Track traders and segments continuously without re-querying manually.
- Curated watchlists for traders, cohorts, and themes
- Live updates integrated with flow/momentum signals
- Watchlist-driven alert policies and routing
Strategy lifecycle state machine
Run each strategy through explicit states so agents don’t “half-deploy” and drift.
- Discrete lifecycle statuses: setup → running → paused → closing
- Safe transitions with guardrails on state changes
- Operator notes attached to lifecycle events for continuity
Real-time strategy introspection
Give agents the operational truth: what’s open, what’s blocked, what the venue sees.
- Open orders visibility tied to strategy instances
- Trading limits/constraints checks before failures happen
- Clearinghouse-aware snapshots for accurate execution context
Manual execution desk for custom agents
For agents trading first-party logic (not mirroring): direct control over positions.
- Open/scale/close positions with explicit direction, leverage, sizing
- Partial reductions without flattening the whole book
- One-call bootstrap: create strategy + deploy initial positions
Position-level order maintenance
Keep risk orders correct as exposure changes—especially during scale-in/out.
- Amend/replace SL/TP as leverage or size changes
- Targeted order cancellation and cleanup primitives
- Execution state sync to avoid orphan orders
Ops health & reliability telemetry
Treat agent trading like production infra: measure reliability, not just PnL.
- Success/error rates and latency statistics by tool/strategy
- Filtering by time range, status, action type for debugging
- Health summaries that support auto-pause and incident response
Incentives, points & leaderboard intelligence
Let agents track rewards context and ranking dynamics programmatically.
- Per-user points breakdown (agent vs perp split, multiplier, rank)
- Leaderboard pagination for efficient monitoring
- Rank movement tracking over time
Cross-market hedging & mispricing capture
Trade one thesis across venues to hedge risk or extract mispricing—perps for delta, prediction markets for event probability.
- Cross-venue position builder (e.g., long perp + opposite event hedge)
- Mispricing detector (implied probability vs market pricing)
- Unified risk view across venues (exposure, max loss, scenario outcomes)
Correlation-aware portfolio risk
Risk management that understands correlated exposure, not just per-asset limits.
- Correlation clustering (BTC-beta, ETH-beta, majors vs alts)
- Portfolio concentration checks by factor, not only by symbol
- Risk budgeting across strategies (allocate “risk units”)
Execution scheduler / order slicing
Improve fill quality with timing and pacing—especially for larger accounts and volatile markets.
- Adaptive order slicing modes (pacing rules, passive-first behavior)
- Cancel/replace logic based on spread and microstructure changes
- “Do-not-chase” rules to avoid adverse selection
Data quality firewall (anti-garbage-in)
Agents should distrust stale or abnormal data before acting.
- Staleness / divergence checks (mark vs oracle vs last trade)
- Abnormal spread/liquidity anomaly detection
- Auto-fallback: degrade to read-only or pause on bad data
Multi-agent orchestration
Separate planning from execution so you can harden safety and reduce failure modes.
- Planner generates trade plan + constraints
- Executor places orders only within policy
- Sentinel monitors risk/telemetry and can halt or reduce exposure
Human-in-the-loop approvals
Let humans intervene only when it matters—without breaking automation.
- Approval thresholds (size, leverage, new market, drawdown breach)
- One-click approve/deny with full context and proposed plan
- Audit trail of approvals, overrides, and emergency actions
Strategy scorecards & benchmarking
Compare strategies honestly using consistent assumptions—not cherry-picked backtests.
- Standard scorecard: drawdown, expectancy, tail loss, stability
- Normalized assumptions (fees/slippage) for fair comparisons
- Leaderboards by regime and timeframe (what works when)
Strategy module marketplace
Let builders ship reusable components and let agents assemble strategies faster.
- Reusable signal modules + risk overlays as composable blocks
- Versioned modules with changelogs and compatibility rules
- Permissioned distribution (private/team/public/paid)
Incident management + post-mortem loop
When something goes wrong, capture evidence automatically and feed learning back.
- Incident triggers (PnL shock, slippage spike, API errors)
- Auto-generated incident pack (logs, orders, state snapshots)
- Post-mortem templates that convert failures into playbooks
Deterministic “dry-run” mode
Run the entire system end-to-end without placing orders—ideal for validation and audits.
- Generate full decision graph + order plan with execution disabled
- Compare planned vs actual once live is enabled
- Hard safety toggles preventing accidental trade placement
Permissioned wallet security & policy keys
Treat execution like a privileged capability with strict boundaries and limits.
- Role-based permissions (read-only / execute / withdraw)
- Spend limits + time-bound keys + strategy-scoped wallets
- Multi-wallet separation (research vs execution vs treasury)
Capital allocator (multi-strategy portfolio manager)
Agents don’t just trade—they allocate risk to what’s working under constraints.
- Allocate capital based on scorecards, regimes, and risk budgets
- Auto-rebalance on schedule or trigger conditions
- Drawdown-aware throttling (reduce risk after losses, restore slowly)