io.github.nirholas/3D-AI-Agent-Avatar

What we checked

Six static checks, weighted by risk. Every result reflects only what could be observed in the published package and repository — never intent.

• Injection surface

  unscannable25% of grade

  Tool descriptions/manifest scanned for instruction-injection patterns (imperatives at the model, hidden text, 'ignore previous', data-exfil URLs).

  No tool descriptions, server instructions, prompts, or resources were fetched; nothing to scan for injection.

• Supply chain

  warn25% of grade

  Package provenance: namespace verification, repo linkage, maintainer count, account age, postinstall scripts, typosquat distance.

  dependency "@gltf-transform/core" is edit-distance 1 from popular package "cors" (typosquat candidate)

• Credential hygiene

  warn15% of grade

  How the server takes secrets (env vs plaintext config vs hardcoded); secrets appearing in tool schemas.

  .env.example: embeds a PRIVATE KEY block. (in an example/sample file — lower severity, but a real key here still leaks.) Reads secrets from the environment (e.g. src/config.js) — the recommended intake shape.

• Permission scope

  pass15% of grade

  Declared tools vs. breadth (filesystem, network, exec); flags shell-exec and unbounded filesystem access.

  Network egress present (src/lib/atomic-launch.js: outbound HTTP/socket call) — expected for most servers and not a finding on its own. No shell-exec, unbounded/sensitive filesystem, or download-and-run primitives found in scanned source. PASS = no static red flag. Static scanning is blind to code fetched or executed at runtime.

• Version behavior

  fail10% of grade

  Diff of tool definitions between versions; new permissions or changed descriptions in a patch release (the postmark-mcp class).

  vs prior 1.2.0 (patch bump): new capability primitive(s) in source: http-client; new hardcoded outbound host/BCC destination(s): aistudio.google.com., api.aixbt.tech, api.avaturn.me, api.cdp.coinbase.com, api.coin-communities.xyz, api.dexscreener.com, api.eu.axiom.co, api.mainnet-beta.solana.com, api.openai.com, api.replicate.com, api.solana.com, api.web3.storage, app.pinata.cloud, arb1.arbitrum.io, avaturn.me, birdeye.so, build.nvidia.com, cdp.coinbase.com, cloud.debank.com., cloud.ibm.com, cloud.sambanova.ai, cloud.walletconnect.com., cloudflare-eth.com, coin-communities.xyz, console.anthropic.com, console.cloud.google.com, console.groq.com, console.mistral.ai, console.upstash.com, console.x.ai, dash.cloudflare.com, dashboard.alchemy.com., dashboard.nodereal.io, dashboard.privy.io, dashscope-intl.console.aliyun.com, dev.helius.xyz, dev.to, developer.twitter.com, developer.x.com, docs.z.ai, drpc.org, editor.avaturn.me, explorer.jito.wtf, facilitator.payai.network, frontend-api-v3.pump.fun, github.com, integrate.api.nvidia.com, lite-api.jup.ag, livekit.io, mainnet.base.org, mainnet.block-engine.jito.wtf, mainnet.helius-rpc.com, medium.com, modelscope.cn, neynar.com., nft.storage., openrouter.ai, owlracle.info, platform.openai.com, platform.solana.com, pub-, pump.fun, quicknode.com, replicate.com, resend.com, rpc.sepolia.org, sepolia-rollup.arbitrum.io, sepolia.base.org, sepolia.optimism.io, siliconflow.com, sketchfab.com, sns-api.bonfida.com, solscan.io, t.me, three.ws, three.ws., us-south.ml.cloud.ibm.com, web3.storage, www.ankr.com, www.blocknative.com, www.bnbchain.org, your-domain.com, z.ai, zauth.inc.

• Transport config

  warn10% of grade

  Remote servers: TLS and auth mode (none/token/OAuth). Local servers: whether the manifest indicates it phones home.

  manifest/source references analytics behavior — disclosure