Robinhood Chain · 4663 · AI agent, in your terminal

The AI agent for Robinhood Chain.

It watches contracts, prepares your transactions to sign, and finds the alpha, all from your terminal.

no private key · no telemetry · never says "safe"

$npm i -g aleph-cli Get Aleph GitHub
$ALEPH 0x3e21311b34346B19b559DfdcECBB3D0Df777C86e virtuals

verify it yourself: aleph scan 0x3e21…C86e

aleph scan 0x…MATE
  ALEPH   MATE (MATE)                     robinhood chain · 4663

  WARNINGS                                    3 findings

  warn  contract not verified     source not published
  warn  ownership not renounced   owner() = 0x9f3a…c21b
  warn  mint capability           mint(address,uint256)
  skip  sell simulation not run   v4 hook pool

  pool     uniswap v4 · PoolManager holds 1,240,000 MATE
  holders  top-10 hold 41.2% · deployer holds 3.1%

  aleph reports what it can read on-chain.
  absence of findings is not safety.
0
runtime deps
0
telemetry
0
tests green
0
chains to mint
What it reads

Four lenses on a contract, before your money moves.

01

Identity

Bytecode, Blockscout verification, name and symbol, age. Is there even a real, published contract here?

02

Deployer power

Ownership, upgradeable proxy, mint, mutable fee, blacklist, pause. Everything the owner can still do to you.

03

Liquidity

Uniswap v4 pool existence via the PoolManager's token balance. Real depth, never a fabricated number.

04

Distribution

Top-holder concentration, the deployer's own share, and the launchpad the token came from.

The honest part

On honeypots, Aleph would rather say nothing than guess.

Aleph does not ship a buy-then-sell honeypot verdict on Robinhood Chain. It says so with a skipped finding instead of guessing. The pools are Uniswap v4 with per-token hooks; a naïve simulation would mislabel a normal token, and a wrong "not a honeypot" on a security tool is worse than nothing. Absence of findings is not safety.
How it works

Free to read. Yours to sign. Never your keys.

Claude Code skills

Drive Aleph from natural language.

Thirty-seven Claude Code skills, in five groups. Each one just runs the CLI and relays what it says. Never "safe", never signs, never asks for a key.

01

Read a contract

What its owner can still do to you: verification, ownership, mint authority, mutable fees, blacklists, pause switches, upgradeable proxies, liquidity, holder concentration, and the honest answer on honeypots.
aleph · aleph-checklist · aleph-proxy · aleph-honeypot · aleph-holders · aleph-pool · aleph-deployer · aleph-compare · aleph-batch · aleph-report · aleph-explain

02

Find what is moving

Discovery instead of waiting: the NFT mints active on-chain right now, new launches and buys from the deployers and wallets you follow, and what any address is holding.
aleph-mints · aleph-alpha · aleph-watchlist · aleph-whale · aleph-wallet

03

Prepare a transaction

Aleph builds it, you sign it. Mints on three chains, liquidity steps, and allowlist registration, all handed back as to / value / data for your own wallet.
aleph-mint · aleph-nft · aleph-lp · aleph-wl

04

Automate and stay covered

A local executor on a burner key you control, a scan gate that refuses flagged contracts, a kill-switch that trips on a rug signal, and hosted alerts that keep running with your laptop closed.
aleph-agent · aleph-snipe · aleph-watchtower · aleph-alerts · aleph-watch · aleph-webhook · aleph-alert-triage · aleph-sentry · aleph-portfolio-watch

05

Access and plumbing

Sign in with a signature, see what is free and what costs credits, point Aleph at your own RPC, wire it into CI, or expose it as MCP tools for another agent.
aleph-account · aleph-credits · aleph-tiers · aleph-premium · aleph-config · aleph-ci · aleph-mcp · aleph-security

every skill is a plain markdown file you can read: github.com/Aleph-Agent/aleph/skills

More you can do with it

One tool, more surfaces than a terminal.

The parts that are not a command you type: what runs on its own, what it plugs into, and what you can verify yourself.

Run it autonomously

A separate local executor signs and broadcasts from a burner key you control, behind a scan gate, an allowlist, and per-tx, cumulative, gas, and expiry caps. It never uploads a key.

Get a kill-switch

The watchtower re-reads your contracts for ownership, pause, and proxy-implementation changes, and freezes the agent the moment one fires.

Plug into other agents

An MCP server exposes scan, alpha, tiers, and mint-prepare as tools for any MCP client. Read-only and prepare-only by construction: no signing tool exists.

Script it in CI

Stable, versioned --json plus exit codes (0 clean, 1 warnings, 2 critical, 3 could not run) let you gate a deploy or a buy on real on-chain state.

Bring your own RPC

The default is the public endpoint and the RPC is yours to change. Nothing is proxied through us, and no scanned address is ever reported anywhere.

Audit it yourself

MIT, open source, zero runtime dependencies, with an in-house keccak256 and a test suite that fails the build if any output claims a token is "safe".

One line to start

Put an agent on the chain.
Keep your keys.

$npm i -g aleph-cli Get Aleph See tiers