Download for MacApple silicon, macOS 14+
Skill

dex-analysis

@allium-labs Updated 2026-10-23

Methodology playbook for analyzing DEX/AMM trade, liquidity, and pricing data on Allium — Uniswap, PancakeSwap, PumpSwap/Pump.fun, Orca, Pendle, aggregators (0x, 1inch), and perp DEXs like Hyperliquid. Use whenever a question touches swap volume, trade counts, liquidity positions, impermanent loss, aggregator fee/routing analysis, or DEX-derived token prices — even if the user just names a protocol or asks why a number "looks off" compared to a known reference. Covers the per-hop double-counting, protocol-vs-project grouping, and fee-leg reconstruction pitfalls that come up repeatedly before you write the SQL.

Install

With your agent

Install OMGSkills skill "allium-labs/allium-claude-plugin:skills/dex-analysis" into this agent's normal skills directory: call the OMGSkills MCP get_install_plan tool for your agent and run the script it returns (it installs only the pinned commit and verifies SKILL.md), and stop if MCP or the pin is unavailable. Do not overwrite or change existing skills without asking me.

Manual command

git clone https://github.com/allium-labs/allium-claude-plugin /tmp/allium-claude-plugin && ln -s /tmp/allium-claude-plugin/skills/dex-analysis ~/.claude/skills/dex-analysis

From README

DEX Analysis The most common DEX-analysis mistake isn't a wrong table — it's treating a per-hop or per-side data model as if it were per-swap or per-trade. Check the grain of the table before aggregating. Trade rows are often per-hop, not per-swap Multi-hop routes (A→B→C) can produce 2-3 trade rows for what the user experienced as one swap. This inflates both volume and row counts — one real comparison found 1.3x inflation vs. an external source on a specific chain purely from hop fan-out. Compare COUNT() vs. COUNT(DISTINCT transactionhash) on a trades table before trusting a volume/count total — a large gap between them is the signature of multi-hop fan-out. When you need to compare volume against an external aggregator (DefiLlama, etc.), either divide by the observed hop ratio or aggregate at transactionhash grain first.