Metis RWA Research

io.github.metis-rwaPublished by GitHub account metis-rwav0.1.1Updated Oct 24, 2026

Research tokenized real-world assets on Solana: issuer controls, supply, liquidity, premiums

VerifiedStreamable HTTPWeb executableFinanceData & Analytics

Overview

AI-generated overview

Metis RWA Research lets an assistant research tokenized real-world assets on Solana: issuer controls, supply, holders, liquidity and premiums.

What you can do
List and filter the built-in registry of tokenized assets by symbol or type, or use a raw Solana mint address instead of a symbol. Read on-chain token state: mint, freeze, seize, pause and transfer-gating controls, supply and caps, holder concentration and frozen balances. Read market state: pool-by-pool liquidity and the premium against the tracked reference security. Run a full observation cycle that compares fresh reads against stored history and reports detections, source conflicts and confidence. Browse published research: list publications, open one with its claims, evidence and methodology, see what is being investigated and check source health. Use the bundled due-diligence and price
When to use it
Use it when you need to answer who controls a tokenized equity or fund on Solana and what that control means for a holder, or to check supply, concentration, liquidity and premium before forming a view. The hosted endpoint suits one-off questions; the local package suits repeated observation with your own RPC endpoint and history.
Requirements
Either the hosted HTTP endpoint (no key, read-only, no session) or the local npm package run over stdio with Node.js. Optional environment variables: SOLANA_RPC_URL (a dedicated endpoint is recommended), METIS_API_URL, METIS_ASSETS_FILE, METIS_MCP_STATE_DIR and METIS_MCP_TIMEOUT_MS. Network access to Solana, DEX venues and a price service.
Before you install
Read-only research, but it is informational and can be wrong, and is not financial advice. A wrong mint address in METIS_ASSETS_FILE returns confident wrong answers, so verify it on chain. Largest token accounts are accounts, not beneficial owners. DexScreener refuses datacenter traffic, so pool data degrades on cloud hosts. Reference prices can be stale outside market hours and look like a premium. Observation history is stored locally by default.

Installation

In SourceWeft

  1. Open Metis RWA Research in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.metis-rwa%2Fmcp/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "metis-rwa-research": {
      "type": "http",
      "url": "https://metisagent.co/mcp"
    }
  }
}

README

Metis MCP

An MCP server for researching tokenized real-world assets on Solana.

It gives an assistant twelve tools: six that read chain and market state live, and six that reach the research the Metis agent has already published. Every figure it returns comes from a fresh read, stamped with its source, the time it was taken, and a hash of the raw payload.

The point is the first question anyone should ask about a tokenized asset and almost nobody can answer quickly: who controls this token, and what happens to my position if they use that control?

Three ways to connect

As a plugin. The easiest path, and the only one that also carries the skills. In Claude Code:

/plugin marketplace add metis-rwa/mcp
/plugin install metis@metis-rwa

In Claude Desktop or on the web, open Customize, then Plugins, then add from a repository with https://github.com/metis-rwa/mcp.git. The plugin connects to the hosted endpoint and adds /metis:due-diligence, /metis:premium-watch, and a token-safety skill that runs whenever someone asks who controls a token. See plugin/README.md.

Hosted, nothing to install. The same tools are served over HTTP at https://metisagent.co/mcp:

bash
claude mcp add --transport http metis https://metisagent.co/mcp

The hosted endpoint answers observation questions from the research agent's own stored history, which goes back further than any single machine can. It reads only, needs no key, and holds no session.

Local, over stdio. Run the package when you want your own RPC endpoint, your own asset list, or history kept on your machine:

bash
npx -y @metisagent/mcp

Or clone and build:

bash
git clone https://github.com/metis-rwa/mcp.gitcd mcpnpm installnpm run buildnode dist/index.js

The server speaks MCP over stdio.

Connect the local server

Claude Code:

bash
claude mcp add metis -- npx -y @metisagent/mcp

Claude Desktop, in claude_desktop_config.json:

json
{  "mcpServers": {    "metis": {      "command": "npx",      "args": ["-y", "@metisagent/mcp"],      "env": {        "SOLANA_RPC_URL": "https://your-endpoint.example/rpc"      }    }  }}

Any other MCP client works the same way: run metis-mcp (or node dist/index.js) and speak MCP on stdin and stdout.

Tools

On-chain checks:

ToolWhat it answers
list_rwa_assetsWhich assets this server knows by symbol, and their mints
get_token_controlsCan the issuer mint, freeze, seize, pause, or gate transfers
get_token_supplyHow many tokens exist, and whether supply is capped
get_holder_concentrationWho holds the supply, and is any of it frozen
get_token_marketHow deep the liquidity is, pool by pool
get_reference_premiumIs the token trading away from the security it tracks

Research:

ToolWhat it answers
observe_assetOne full research cycle: read everything, compare against history, report detections and confidence
get_observation_historyWhat this server has recorded for the asset so far, and the trend
list_researchWhat the Metis agent has published, filtered by asset or category
get_researchThe full object behind one publication: claims, evidence, methodology, provenance
get_research_statusWhat the agent is investigating right now, and source health
check_sourcesWhich data sources are answering, and how fast

Every tool takes either a registry symbol (TSLAx) or a raw Solana mint address, so assets outside the registry work too.

The registry ships with the 72 xStocks on Solana, from AAPLx through XOMx, including the index and commodity fund shares (SPYx, QQQx, GLDx, PPLTx, TBLLx) and the private-market listings (SPCXx, VCXx). Filter it with list_rwa_assets, by substring or by asset type.

Two prompts ship with the server: rwa_due_diligence walks an asset through controls, supply, concentration, depth, and premium, and premium_watch reads a price gap against stored history. Two resources are exposed as well: metis://registry and metis://methodology.

What get_token_controls is for

A tokenized equity is a claim on something held off chain, so the issuer keeps powers a plain token does not have. Token-2022 makes those powers explicit on the mint, and this tool translates each one into what it means for a holder:

  • Permanent delegate: tokens can be moved or burned out of any account without the holder signing.
  • Freeze authority: any account can be frozen, which blocks selling and redemption.
  • Transfer hook: every transfer runs through a program that can reject it, which is how an allowlist is enforced.
  • Default account state frozen: new accounts cannot transact until the issuer thaws them, so the token is permissioned by default.
  • Pausable: transfers can be halted chain-wide.
  • Scaled UI amount: displayed balances are rescaled, which is how splits and similar corporate actions land on chain.

None of these are defects. Redeemable backed assets need most of them. They are simply facts about the asset that belong in any answer about whether it is worth its reference price.

Configuration

VariableDefaultPurpose
SOLANA_RPC_URLa public nodeSolana JSON-RPC endpoint
METIS_API_URLhttps://metisagent.coMetis research API
METIS_ASSETS_FILEnoneJSON file of extra assets to register
METIS_MCP_STATE_DIR~/.metis-mcpWhere observation history is stored. Set to off for memory only
METIS_MCP_TIMEOUT_MS20000Network timeout per request

Set SOLANA_RPC_URL to an endpoint you control if you can. Public nodes throttle the heavier reads, and get_holder_concentration is the first one they refuse.

METIS_ASSETS_FILE takes an array. Only symbol and mint are required:

json
[  {    "symbol": "AAPLx",    "name": "Apple xStock",    "mint": "<mint address>",    "decimals": 8,    "assetType": "tokenized_equity",    "referenceTicker": "AAPL",    "issuerName": "Backed Assets"  }]

Verify a mint against the chain before adding it. A wrong address returns confident, wrong answers. Everything already in the registry was read back from Solana first: the names and symbols in src/assets.ts are the ones written into each mint's own Token-2022 metadata, not labels copied from a token list. Check them again at any time:

bash
npm run verify-assets

How an observation cycle works

observe_asset reads supply and holder concentration from Solana, pooled liquidity and volume from DEX venues, and the token and reference prices from a public price service. It compares the result against the observations already stored for that asset, then reports:

  • Observations: what is true right now, each tied to the reads behind it.
  • Detections: metrics that crossed a threshold, with the baseline they crossed it against.
  • Conflicts: sources that disagree, recorded rather than averaged away.
  • Confidence: computed from source coverage, source failures, cross-source agreement, and how much history backs the comparison. A model's own stated confidence is never an input.

History lives on the machine running the server. The first call on an asset has no baseline, so only the premium threshold can fire. Call it again over time and supply, liquidity, activity, and concentration detections come alive. The hosted endpoint skips that warm-up entirely: it compares against the history the research agent has been recording every ten minutes since it went live.

Limits worth stating

  • Largest token accounts are accounts, not beneficial owners. Pools, bridges, and custodial accounts sit at the top of most lists.
  • DexScreener refuses datacenter traffic. Run the server from a laptop and it answers; run it from a cloud host and pool data falls back to the price service, which reports liquidity, 24h volume, and a holder count but no pool breakdown and no transaction counts. Tools say when a number came from the standby.
  • A republished reference price can be stale outside market hours, and that alone can look like a premium.
  • Aggregated venue data covers listed pools. Anything traded over the counter or internally is invisible here.
  • This is informational research produced by software. It is not financial advice, and it can be wrong.

Development

bash
npm installnpm run typechecknpm testnpm run inspector

Tests cover the analysis rules, the control translations, the registry, and the history store. They run against the build, so npm test builds first.

Releasing

The package publishes to npm as @metisagent/mcp, and server.json describes it for the MCP registry. The registry proves ownership by matching mcpName in package.json against the server name in server.json, so those two strings have to stay in step.

  1. Bump the version in package.json and in both places in server.json, then add a CHANGELOG.md entry.
  2. npm run verify-assets to confirm the registry still matches the chain.
  3. npm test.
  4. npm publish --access public.
  5. mcp-publisher login github then mcp-publisher publish to list the release in the MCP registry.

License

MIT

Source: README.md at commit c3a0cb9

Tools

0
SourceWeft has not read this server's tools yet.

Version history

1
  1. v0.1.1LatestSep 16, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declaredRequired10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts