Research tools / Built for your agent

Research with
an edge.

Web3 research skills for Claude and Codex.

Explore skills ↗
Built forClaudeCodex

Give your agent a specialty.

Less noise. More context.

02 / The skill library

Find your edge.

One skill. A new way to read the chain.

Methods, tools, sources. Packaged.

How it works

From question to clarity.

01

Choose your lens.

Pick a focused research angle, like a protocol, narrative, or onchain activity.

02

Give it to your agent.

Turn your question and lens into a repeatable research skill for Codex or Claude.

03

Follow the evidence.

Get a sourced brief with clear findings, assumptions, and links to the original sources.

> Research brief

question:
Is EigenLayer’s restaking model sustainable long term?
lens:
protocol fundamentals
scope:
mechanisms, demand, risks
—

Under the hood

A skill.
Unpacked.

Inspect the method behind the answer.

# Research method

1. Define the question
2. Collect primary evidence
3. Check every claim
4. Write a sourced brief

A repeatable research method.

Clear steps. Explicit assumptions. Evidence for every finding.

Explore the documentation ↗

Give your agent a new perspective

Read the chain.
Find the signal.

Explore skills ↗
Illustrative report / Sample data

Wallet research brief

An example of the structure. Not a live wallet analysis.

NETWORK / SOLANA
ADDRESS / DEMO WALLET
WINDOW / 30 DAYS · ILLUSTRATIVE

01 / Observations

The supplied example dataset contains swaps, token transfers and protocol deposits. A real report would link each material finding to a transaction or source.

02 / Interpretation

Repeated interactions may suggest a recurring workflow. They do not establish the owner’s identity or prove that a transfer was a sale.

03 / Missing context

Cost basis, offchain activity and ownership labels are unknown. No return estimate is provided without sufficient inputs.

04 / Evidence checklist

Transaction links · Source timestamps · Token identifiers · Coverage limitations