An agentic skills framework & software development methodology that works.
Trading skills pack for Claude Code and agents
Superior Skills is a bundle of trading strategy skills and tool schemas for Superior Trade. The entry skill walks an agent through account setup, funding, strategy selection, backtesting, and live deployment, then hands off to venue-specific skills like Hyperliquid or Polymarket. The repo also includes reusable strategy templates, regime filters, exit logic, and validation scripts that keep the skills aligned with the Unified API.
Builders who want an agent to author, test, and run trading strategies with reusable skills instead of one-off prompts.
You can move from a new account to a running strategy with less manual prompting and more reusable structure.
What it does
Entry skill for the full trading flow
`skills/superior-trade/SKILL.md` covers the path from API key and funding to strategy choice, backtest, deployment, and monitoring.
Strategy skills library
`skills/` includes templates and focused skills such as `backtesting`, `trade-thesis`, `regime-overlay`, `dsl-exit-engine`, and venue skills like `hyperliquid` and `polymarket`.
Unified API references
`references/unified-runtime.md` and the API comparison table define the current request and resource model the skills should use.
Validation and sync scripts
`scripts/validate.mjs`, `scripts/validate-skills.mjs`, and related checks help keep skill files and endpoint claims consistent.
Plugin packaging
`.claude-plugin/plugin.json` and `.claude-plugin/marketplace.json` package the repo as an installable Claude-style plugin.
README
Superior Skills
Trading strategies and intelligence tools for Superior Trade — natural-language strategy authoring, backtesting, and autonomous deployment on Hyperliquid, Lighter, and Polymarket.
Designed for OpenClaw users adding trading capabilities to their agent, and for traders who want validated templates rather than rolling their own.
Start here
superior-trade is the entry skill. It carries the whole path — API key, trading account, funding, strategy selection, backtest, live deployment, monitoring — and hands off to the venue skill once the user has picked one. Point an agent at this skill and it can take someone from no account to a running strategy without further instruction.
Everything below is what it routes to.
Unified API migration
All Superior Trade operations in this package use the unversioned Unified API.
The committed OpenAPI snapshot and endpoint audit reject versioned API guidance.
Read references/unified-runtime.md for the
shared lifecycle.
Use this single prompt to migrate another skill:
Migrate this Superior Trade skill to the Unified API. In the Superior skills package, read references/unified-runtime.md and scripts/api-contract/unified.json first. Replace every versioned Superior Trade endpoint, method, payload, polling step, and response assumption with the functionally equivalent Unified API operation. Do not keep compatibility fallbacks. When Unified API has no equivalent, state that the operation is unavailable instead of inventing a route. Preserve the skill's domain strategy and safety rules, update examples and linked references, then run pnpm run validate and fix every endpoint-audit error.
Operation comparison
Equivalent operations across the previous API surfaces are grouped together. "Unavailable" means the skill must report the limitation instead of falling back to a versioned route. See the full migration guide for the product-level lifecycle.
| Operation | Previous API | Unified API | Change |
|---|---|---|---|
| Health check | GET /health | GET /health | Direct |
| Request API access | POST /auth/sign-in/magic-link | POST /account/register, then POST /account/verify | Replaced by an OTP flow |
| List API keys | GET /auth/api-key | GET /account/keys | Direct |
| Create API key | POST /auth/api-key | POST /account/keys | Direct |
| Delete API key | DELETE /auth/api-key/{id} | DELETE /account/keys/{id} | Direct |
| Rename API key | PATCH /auth/api-key/{id} | — | Unavailable |
| Read account identity | No dedicated equivalent | GET /account | New account, plan, limits, and usage view |
| Read account usage | No dedicated equivalent | GET /account/usage | New |
| Read trading account / managed wallet | GET /v2/account, GET /v3/account, venue balance reads | GET /wallet | Same managed-wallet domain; the collection becomes the authenticated wallet summary |
| Create managed trading wallet | POST /v2/account, POST /v3/account | Implicit on first use | No explicit Unified creation call |
| Rename trading account | PATCH /v2/account/{address}, PATCH /v3/account/{address} | — | Unavailable |
| Read venue availability | GET /v2/account/{address}/status/{exchange}, GET /v3/account/{address}/status/{exchange} | GET /context/venues, GET /wallet | Consolidated capability and wallet-readiness checks |
| Bootstrap a venue account | POST /v3/account/{address}/hyperliquid, /polymarket, or /lighter | Implicit for the managed path when a deployment starts | No public bootstrap mutation |
| List backtests | GET /v1/backtesting, GET /v2/backtesting | GET /runtime/backtests | Direct |
| Create backtest | POST /v1/backtesting, POST /v2/backtesting, POST /v3/backtest | POST /runtime/backtests | Creation automatically queues the run |
| Get backtest | Versioned GET by ID | GET /runtime/backtests/{id} | Direct |
| Read backtest status/result | Versioned /status and /result reads | GET /runtime/backtests/{id} | Embedded in the resource |
| Start backtest | Versioned status update | — | Removed; creation queues the run |
| Update backtest | PATCH /v2/backtesting/{id} | — | Unavailable |
| Delete or cancel backtest | Versioned DELETE by ID | DELETE /runtime/backtests/{id} | Direct |
| Read backtest logs | Versioned GET logs by ID | GET /runtime/backtests/{id}/logs | Direct |
| Check dataset availability | GET /v2/backtesting-data/hyperliquid, /binance, or /aerodrome | GET /context/datasets, GET /runtime/backtests/dataset | Catalog plus exact-market lookup |
| Scan intelligence | GET /v2/intelligence/scan | GET /context/scan | New parameters and response schema |
| Read symbol setup | GET /v2/intelligence/setup/{pair} | GET /context/setup/{symbol} | Renamed identifier and response model |
| Search markets | POST /v3/markets/search | GET /context/markets | Search command becomes a context read |
| Read leaderboard | GET /v2/leaderboard-strategies | GET /context/leaderboard | Direct |
| List/read tracked traders | GET /v2/copy-trading/traders, GET /v2/copy-trading/traders/{wallet} | GET /context/traders, GET /context/traders/{wallet} | Direct |
| Read candles | No dedicated equivalent | GET /context/candles | New |
| Read funding | No dedicated equivalent | GET /context/funding | New |
| List runtime frameworks | No dedicated equivalent | GET /runtime/frameworks | New |
| List deployments | Versioned deployment lists | GET /runtime/deployments | Consolidated |
| Create deployment | Versioned live and paper creation | POST /runtime/deployments | framework, venue, and mode are request fields |
| Get deployment | Versioned GET by ID | GET /runtime/deployments/{id} | Direct |
| Update deployment metadata | PATCH /v2/deployment/{id} | PATCH /runtime/deployments/{id} | Direct |
| Delete deployment | Versioned DELETE by ID | DELETE /runtime/deployments/{id} | Direct |
| Read deployment status | Versioned /status reads | GET /runtime/deployments/{id} | Embedded in the resource |
| Start or stop deployment | Versioned PATCH or PUT status operations | PUT /runtime/deployments/{id}/status | Standardized lifecycle action |
| Attach deployment credentials | Versioned POST credentials operations | PUT /runtime/deployments/{id}/credentials | Method and contract changed |
| Read deployment credentials | GET /v2/deployment/{id}/credentials | — | No standalone credential read; safe metadata may appear on the deployment |
| Read deployment logs | Versioned GET logs by ID | GET /runtime/deployments/{id}/logs | Direct |
| Read deployment metrics | Paper profit and pod-proxy reads | GET /runtime/deployments/{id}/metrics | Normalized across frameworks |
| Paper deployment | /v2/paper-deployment resource tree | POST /runtime/deployments with mode: "paper" | Mode becomes a field, not a separate resource |
| Deployment history | GET /v2/deployment-history | GET /runtime/deployments | No dedicated history resource |
| Exit all positions | Versioned deployment and portfolio /exit operations | — | Unavailable |
| Deposit to a venue | Versioned Hyperliquid, Polymarket, and Lighter deposit operations | Fund the address from GET /wallet; allocation occurs on managed deployment start | Removed as a public venue mutation |
| Read deposit history | Venue reconciliation reads | GET /wallet/deposits | Read-only history, not a transfer action |
| Withdraw Hyperliquid funds | POST /v3/portfolio/hyperliquid/withdraw | POST /wallet/withdraw | Unified, tracked withdrawal to the verified login wallet |
| Withdraw Lighter funds | POST /v3/portfolio/lighter/withdraw | — | Unavailable |
| Read withdrawal history/status | Venue reconciliation reads | GET /wallet/withdrawals, GET /wallet/withdrawals/{id} | Unified wallet withdrawals only |
| Hyperliquid order/cancel | POST /v2/authorize-and-send/hyperliquid | POST /runtime/executions | Typed order and cancel actions only |
| Polymarket order/cancel | POST /v3/authorize-and-send/polymarket | POST /runtime/executions | Typed placeMarketOrder and cancelOrder actions only |
| Lighter signed action | POST /v3/authorize-and-send/lighter | — | Unavailable |
| List/read execution records | No dedicated equivalent | GET /runtime/executions, GET /runtime/executions/{id} | New durable, sanitized records |
| Hyperliquid brackets | /v2/bracket resource tree | — | No bracket-resource equivalent |
| Hyperliquid wallet transfer | POST /v2/portfolio/hyperliquid/transfer | — | Unavailable |
| Discover the HTTP contract | Versioned OpenAPI documents | GET /openapi.json | One authoritative contract |
| Discover or call MCP | No dedicated equivalent | GET /.well-known/mcp.json, POST /mcp | New |
Hyperliquid and Aerodrome
These integrations use Unified runtime resources with the Freqtrade framework.
Validated strategies (with backtest evidence)
These strategies have backtest evidence on real Hyperliquid data. Numbers are full-period (162 days, 2025-11-20 → 2026-05-01) unless noted.
| Strategy | Regime | Pairs tested | Trades | Win | Profit | Max DD | File |
|---|---|---|---|---|---|---|---|
| Donchian Strong-Regime | Strong directional trend | BTC | 6 | 100% | +6.69% | 0% | donchian-strong-regime |
| Bollinger Reverter 4h | Range / chop (ADX<25) | BTC/ETH/SOL/DOGE | 84 | 65.5% | +8.77% | 18.5% | bollinger-reverter-4h |
Paired together as separate sub-accounts, the two strategies are regime-complementary: the Donchian gate fires zero trades during the chop windows where the Bollinger reverter thrives, and the Bollinger reverter mildly underperforms during the strong-trend windows the Donchian captures.
Template strategies (starting points)
Reference templates for adapting to your own thesis. Backtest before deploying.
dca-weekly— dollar-cost averaging with scheduled buysgrid-trading— profit-laddered position adjustmentfunding-rate-arbitrage— negative-funding capture (carry)funding-squeeze— funding-extreme squeeze ridebasis-arb— spot-perp basis convergencebreakout— Donchian breakout with trailing stopmean-reversion— Bollinger band fade (4h validated; see file)scalping— RSI + volume-thrust template
Categories covered: trend-following, mean-reversion, carry, arbitrage, scalping.
Reusable primitives
Building blocks that compose across strategies.
trade-thesis— Structured pre-trade thesis builder: bull/bear cases, invalidation criteria, and sizing rationale before any live deployment of a new strategy idea. Aliases: pre-trade analysis, conviction check, trade plan, bull/bear case.regime-overlay— Triple-confirmation regime gate (EMA separation + ADX + N-bar return). Turns fragile directional strategies into regime-robust ones. Aliases: regime filter, trend gate, directional confirmation.dsl-exit-engine— Three-phase exit primitive: ROI ladder, hard stop, ratcheting trailing stop. Aliases: ratcheting trailing stop, two-phase exit, take-profit ladder.fees-optimizations— Maker (ALO) vs taker (MARKET) order-type decisioning, builder fees, parameter sweeps. Aliases: fee optimizer, ALO vs MARKET, maker pricing.backtesting— Window selection, walk-forward, parameter sweeps.intelligence— Opportunity scanner / pair-ranking system.
Exchanges
- [`hy
Files in the repo
- .claude-plugin
- docs
- references
- scripts
- skills
- .clawhubignore
- .gitignore
- .turbo-upstream.json
- AUTHORING.md
- LICENSE
- package.json
- README.md
- SKILL.md
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More skills

Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.
Topic in, narrated explainer video out. A Claude Code / Codex skill that turns any topic into a black-canvas motion-graphics explainer video with TTS voiceover, subtitles and a chapter progress bar. Chinese or English; every frame drawn in code with Remotion.
Public repository for Agent Skills
Open-source AI job search: scan job portals, evaluate listings into a structured A-H report with a global 1-5 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)

Production-grade engineering skills for AI coding agents.