Research Hyperliquid Traders with Senpi's Agent Skill

Where this fits

A developer is building a copy-trading agent for Hyperliquid. The agent needs to answer "who should I copy?" and "is this trader any good?" without manually stitching together leaderboard data, trader history, and live order books. This skill encapsulates that research workflow, giving the agent a pre-built due-diligence engine that ranks traders by copyability and flags whether a trader's book can actually be mirrored right now.

Why agents benefit

• Rank by copyability, not just PnL – The hidden engine combines track record with live book depth, distance-from-entry, and 4h momentum, so the agent can prioritize traders whose positions are realistically mirrorable.
• Vet a specific trader before mirroring – The skill scores whether a trader's performance is a hot streak or a real edge, using the same signals a human analyst would check.
• Save tokens and API calls – Instead of making multiple discovery calls and interpreting raw data, the agent gets a distilled research report and can focus on judgment calls.
• Make the final decision – The skill explicitly hands the decision back to the agent: "you make the call." It's a research assistant, not an autopilot.
• Works with the Senpi platform – Uses a USER-scoped token, meaning the agent can act on behalf of a specific user's account, not just generic public data.

Practical scenarios

Finding top copy traders for the month

The user asks "best traders this month." The agent invokes the skill, which pulls leaderboard data, filters for traders with strong 4h momentum and healthy distance-from-entry, and returns a ranked list with reasons. The agent then presents the top three with a summary of why they're worth copying.

Vetting a specific trader before mirroring

A user wants to copy 0x… but is unsure. The agent runs the skill on that address. It returns a scorecard: win rate, drawdown, recent momentum, and a copyability score based on the trader's current open positions. The agent explains the risks and lets the user decide.

Comparing two candidates

The user is torn between two traders. The agent runs the skill on both and outputs a side-by-side comparison, highlighting which one has a more mirrorable book and a more consistent track record. This saves the user from manually inspecting each trader's history.

Add it to your agent workflow

Install the skill with:

npx skills add Senpi-ai/senpi-skills --skill senpi-trader-research

Then in your agent's system prompt, you can invoke it like:

Use senpi-trader-research to answer "who should I copy?" or "is this trader good?" — it will return a ranked list of candidates with copyability scores.

No need to wire up separate discovery calls; the skill handles the data gathering and initial scoring.

Compare before adopting

Before committing, compare this skill against building your own research pipeline. Key points: quality signals – Does the hidden engine use up-to-date Hyperliquid data? Check the repo's commit history and star growth. Maintenance freshness – Hyperliquid's API changes; see how often the skill is updated. Alternatives – Are there other skills for copy-trader research? Evaluate whether they offer similar depth. Workflow fit – Does the skill integrate cleanly with your agent's existing toolchain, especially the Senpi token requirement? If you're already on Senpi, this is a natural fit.

Why it is worth tracking

With 119 stars and an Apache-2.0 license, this skill shows community interest in agent-driven copy trading. The fact that it packages judgment (copyability scoring) rather than just raw data makes it a step above typical research tools. As the Senpi ecosystem grows, this skill is likely to evolve with better signals and broader exchange support. It's worth evaluating now if you're building agents that trade on Hyperliquid, and you can learn more at OpenAgentSkill.