gmgn-holder-analysis Skill for Claude: What It Does and How to Install It

Makes Claude run a token holder report: chip spread, whale and dev behavior, risk wallets and a rating.

Installs: 10.9K; on skills.sh
Stars: 636; on GitHub
Files: 1; besides SKILL.md
Updated: Oct 2026; last repo commit

gmgn-holder-analysis makes Claude run a holder report for a token you name. You give a chain and a token address, and a Python script reads the top 100 holders through the gmgn-cli tool. The report covers chip concentration, entry cost, whale, dev and KOL behavior, risk wallets such as rat traders, bundlers and snipers, linked wallets and smart money signals, ending in a rating.

The rating comes only from token structure. Claude pastes the script output verbatim with no commentary. It needs the gmgn-cli tool, Python 3 and a GMGN API key. The report language follows the language you write in, Chinese or English, and SKILL.md defaults to Chinese.

Good for

Traders who want a holder-structure check before buying a token
Researchers looking at whale, dev and KOL behavior
Anyone checking a token for rat traders, bundlers and snipers
Claude Code users who already run GMGN skills

Install gmgn-holder-analysis

Claude Code

Copy this prompt.

Install the gmgn-holder-analysis skill in Claude Code. Run this command:

npx skills add GMGNAI/gmgn-skills --skill gmgn-holder-analysis -g -a claude-code

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

Pick one route. The --skill route installs only this skill, with -g so it lands in your user folder where SKILL.md expects it. The README's own command (labelled all GMGN skills) installs the whole set; the marketplace manifest describes 12, and the swap skill needs a private key that this skill does not use. SKILL.md runs analyze.py from ~/.claude/skills/gmgn-holder-analysis/, so install it where Claude Code reads personal skills. Every route needs the gmgn-cli tool, Python 3 and a GMGN API key.

From the vendor's docs. Source.

What's in the skill

When it loads

Token holder chip analysis — deep analysis of holder structure including chip distribution, entry cost, whale/dev/KOL behavior, risk wallets (rat traders, bundlers, snipers), related wallets, smart money signals, and an AI rating based purely on token structure. Use when user asks about holder analysis, 筹码分析, 持仓分析, chip structure, who is holding, or whether a token is safe to buy based on its holder composition.

Sections in SKILL.md

Analysis Script
Output Rule
Field Reference
"Zero cost" must be proven, not inferred from a missing buy count
The token's own contract address is not a holder wallet
Holder object key fields
Created-tokens response fields
Rating Standard
Per-metric flag thresholds
Chip quality (footer)
Supported Chains
Dev lookup is dual-source
Notes

Key facts

Author: GMGN (marketplace manifest owner)
License: MIT
Installs on skills.sh: 10,942 (Nov 2, 2026)
GitHub stars: 636 for the whole gmgnai/gmgn-skills repo (Nov 2, 2026)
Last repo commit: Oct 29, 2026
What it includes: 2 files: SKILL.md and analyze.py (a Python script)
Needs: The gmgn-cli tool (npm), python3 and a GMGN API key from gmgn.ai/ai
Chains: sol, bsc, base, eth, robinhood, arc, stable
Part of: gmgn-skills. The marketplace manifest describes the repo as 12 GMGN skills; this one is the holder chip analysis
Works in: SKILL.md hard-codes ~/.claude/skills/gmgn-holder-analysis/analyze.py, so it is written for Claude Code. The README also documents Codex for the repo's skills

What it includes

Runs gmgn-cli config --check before any command
If it fails, SKILL.md tells Claude to run gmgn-cli config and apply the key with gmgn-cli config --apply. If gmgn-cli is unknown it points you to npm install -g gmgn-cli.
Runs code · Config check
python3 analyze.py <address> <chain> <lang>
The script calls gmgn-cli for token holders, a dev lookup, native token price and the creator's token history.
Runs code · analyze.py
sol for Solana; auto for 0x addresses unless you name a chain
With auto, the script tries bsc, eth and base in order and falls back to eth. Supported chains are sol, bsc, base, eth, robinhood, arc and stable.
Read · Chain choice
Shares are of float, not total supply
Float is 1 minus burn minus DEX. Only the top 100 holders are fetched, so a percentage is a floor unless those 100 cover all of the float.
Read · Tradeable-float percentages
Cannot Assess, Not Recommended, Caution, Light Position, Normal
Not Recommended fires on rat traders over 5%, a wallet over 10%, or a dev sock puppet. Caution needs two or more warnings, Light Position exactly one.
Read · Rating
Top10, Top20, airdropped chips, risk wallets, linked funding, zero balances
Each has red, yellow and green thresholds. Diamond hands invert, where more is better.
Read · Per-metric flags
No percentages when float is under 2% or the holder list is empty
The script prints a banner, shows n/a and rates Cannot Assess instead of reporting no risk.
Read · Guards against bad data
Primary dev tag lookup with a token info fallback
Adds the creator's token history and all-time high market cap, and warns when that high looks stale.
Network · Dev and creator lookup
Paste the script output verbatim, with no extra text
SKILL.md says every line and section, nothing summarized, with no introduction or commentary.
Read · Output rule

From the official docs

Permissions

Medium risk

What it can do

Read your data
Create and change things
Delete things
Read-only mode

Watch out for

Claude may apply the key for you
If gmgn-cli config --check fails, SKILL.md tells Claude to run gmgn-cli config and then apply the key with gmgn-cli config --apply <KEY>. Create ~/.config/gmgn/.env yourself first so that fallback is not needed. github.com
The same key file can unlock trading skills
The repo's swap and order commands need GMGN_PRIVATE_KEY in the same ~/.config/gmgn/.env, and installing the whole repo adds those skills. Leave the private key out unless you want trading. github.com
It runs a bundled Python script
analyze.py is 56 KB and calls gmgn-cli many times in parallel threads. Read it before first use. github.com
It sends the token and chain to GMGN
The address and chain you give are passed to gmgn-cli, which queries GMGN's API with your key. github.com

Security advisories

No published security advisories as of Nov 2, 2026.

Technical details

Turning on read-only	No read-only mode. The script calls token holders, token info and portfolio created-tokens.
Talks to	GMGN API, reached through gmgn-cli (the token address, chain and wallets in the report)
Sign-in and keys	README: GMGN_API_KEY is written to ~/.config/gmgn/.env (a project .env takes precedence), with chmod 600 advised. The same file can hold GMGN_PRIVATE_KEY.
Scopes	SKILL.md runs python3 on analyze.py, which runs gmgn-cli through subprocess gmgn-cli calls GMGN's OpenAPI with your API key SKILL.md may run gmgn-cli config --apply to save a key

Questions

What does the gmgn-holder-analysis skill do?

It makes Claude run a holder report for a token. The report shows how the tradeable supply is spread across wallets, what whales, the dev and KOLs are doing, which wallets look risky, and a rating based only on token structure.

How do I install gmgn-holder-analysis?

Use the install box at the top of this page. It needs the skill, the gmgn-cli tool and a GMGN API key. SKILL.md runs the script from ~/.claude/skills/gmgn-holder-analysis/, so use the Claude Code tab.

Do I need an API key?

Yes. The README says to create your own key at gmgn.ai/ai for real use. It also mentions a public demo key for testing the CLI. The swap commands use an extra private key, and this skill's script does not call them.

Does it need Python?

Yes. SKILL.md runs python3 on analyze.py, and the script imports only standard library modules.

Is the rating financial advice?

No. SKILL.md describes the rating as based purely on token structure. It does not look at price targets or news, so treat it as one input.

Why do my results show n/a?

SKILL.md says that when under 2% of supply is tradeable, or the holder list comes back empty, the script prints no float percentages and rates the token Cannot Assess. It does this instead of showing numbers that look like findings.

Why do I get 401 or 403 errors?

The README says gmgn-cli commands only work over IPv4, and that errors with correct credentials can mean your traffic is going over IPv6.

Is it safe?

Installing it copies a SKILL.md and a Python script. Using it makes Claude run that script, which calls gmgn-cli, so the token address and chain you give go to GMGN through your API key. Keep the key in a file only you can read, as the README does with chmod 600, and keep your swap private key out of the same file unless you want the trading skills to work.

Information

updated Nov 2, 2026

Repository: gmgnai/gmgn-skills
Folder: skills/gmgn-holder-analysis

gmgnai/gmgn-skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 6

gmgn-holder-analysis SKILL.md · github.com
gmgn-skills README · github.com
.env.example · github.com
skills CLI commands · github.com
analyze.py · github.com
gmgn-cli package.json · github.com