registry / @threenative/core / 0.3.6

@threenative/core@0.3.6

The small vanilla Three.js game runtime foundation.

AI Security Review

scanned 2d ago · by lpm-firewall-ai

The postinstall hook performs first-party MCP extension setup in supported project configurations, with a target check and opt-out. This leaves a concrete but guarded agent extension lifecycle risk, so a warning is appropriate.

Static reason

High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review

Trigger

Installing the package when the target project declares it, unless the setup skip variable is set.

Impact

An agent host may start the package’s tools in the project context without a separate setup command.

Mechanism

The installer writes local Node commands for the package’s MCP servers into project-level host configuration files.

AI rationale

Evidence

package.jsonscripts/postinstall.mjsscripts/ensure-mcp.mjsmcp/install.mjsmcp/servers.mjs.mcp.json.codex/config.toml.cursor/mcp.json.vscode/mcp.json.gemini/settings.jsonopencode.json.zed/settings.json

Decision evidence

public snapshot

AI called this Suspicious at 86.0% confidence as Dangerous Capability with medium false-positive risk.

Evidence for warning

The manifest runs a postinstall hook, which calls MCP setup automatically.
MCP setup runs unless its skip variable is set to “1” or “true”.
The installer targets a project that declares this package and writes agent configuration files there.
The configuration entries launch local Node processes for package-owned MCP tools.
The Blender recipe path is built from the packaged scripts directory and read when a recipe is requested through the MCP tool.

Evidence against

The installer skips nested node_modules targets and the package’s own workspace, and requires a project manifest.
The setup writes project-level host configuration; the inspected Blender operation reads a packaged recipe rather than establishing an unrelated network or account action.

Behavioral surface

Source

ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetwork

Supply chain

HighEntropyStrings

Source & flagged code

9 flagged

package.jsonView file

•scripts.postinstall = node ./scripts/postinstall.mjs

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•scripts.postinstall = node ./scripts/postinstall.mjs

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

mcp/launch.mjsView file

•matchType = normalized_sha256 matchedPackage = @threenative/core@0.3.4 matchedPath = mcp/launch.mjs matchedIdentity = npm:QHRocmVlbmF0aXZlL2NvcmU:0.3.4 similarity = 1.000 summary = normalized source hash matched finalized malicious source

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

mcp/launch.mjsView on unpkg

7* here: `threenative-engine-mcp` publishes an `exports` map with only an `import` condition, which L8: * `createRequire().resolve` refuses, and the answer we want is the directory rather than a L9: * specifier's entry point. Walking `node_modules` upward from both the project and this package

Medium

Dynamic Require

Package source references dynamic require/import behavior.

mcp/launch.mjsView on unpkg · L7

scripts/apply-three-patch.mjsView file

18options.consumerRoot ?? L19: process.env.INIT_CWD ?? L20: process.env.npm[redacted] ?? L21: process.cwd(), L22: ); L23: const threeRoot = options.threeRoot ?? resolveThreeRoot(consumerRoot, packageRoot); L24: const manifestPath = path.join(threeRoot, "package.json"); L25: const manifest = await readJson(manifestPath, "Three.js package manifest"); ... L53: const resolver = createRequire(path.join(root, "package.json")); L54: const entry = resolver.resolve("three"); L55: const resolvedRoot = findPackageRoot(entry); ... L110: try {

Low

Weak Crypto

Package source references weak cryptographic algorithms.

scripts/apply-three-patch.mjsView on unpkg · L18

mcp/install.mjsView file

1Install-time code directly mutates a foreign AI-agent control surface: L1: import { existsSync, mkdirSync, readFileSync, writeFileSync } from "node:fs"; L2: import path from "node:path"; L3: import { CODEX_MCP_SERVERS, mergeMcpServers } from "./servers.mjs"; L4: ... L45: label: "Claude Code", L46: file: ".mcp.json", L47: format: "mcpServers", L48: }), L49: Object.freeze({ id: "codex", label: "Codex", file: ".codex/config.toml", format: "codex" }), L50: Object.freeze({ id: "cursor", label: "Cursor", file: ".cursor/mcp.json", format: "mcpServers" }), L51: Object.freeze({ id: "vscode", label: "VS Code", file: ".vscode/mcp.json", format: "vscode" }), ... L55: file: ".gemini/settings.json", Write operation from mcp/install.mjs: L1: import { existsSync, mkdirSync, readFileSync, writeFileSync }

Critical

Ai Agent Control Hijack

Source creates an unconsented AI-agent control surface through install-time mutation or a default unauthenticated remote skill channel.

mcp/install.mjsView on unpkg · L1

gpl/convert.pyView file

•path = gpl/convert.py kind = build_helper sizeBytes = 6997 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

gpl/convert.pyView on unpkg

gpl/fixtures/make_world_fixture.pyView file

•path = gpl/fixtures/make_world_fixture.py kind = payload_in_excluded_dir sizeBytes = 7323 magicHex = [redacted]

High

Payload In Excluded Dir

Package hides binary, compressed, or executable-looking payloads in test/fixture/hidden paths.

gpl/fixtures/make_world_fixture.pyView on unpkg

mcp/blender-server.mjs#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 3

High

Semantic Analysis Limited

A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.

mcp/blender-server.mjs#virtual:normalized:round1View on unpkg

Source & flagged code

9 flagged · loading source

package.jsonView file

•scripts.postinstall = node ./scripts/postinstall.mjs

High

Install Time Lifecycle Scripts

Package defines install-time lifecycle scripts.

package.jsonView on unpkg

•scripts.postinstall = node ./scripts/postinstall.mjs

Medium

Ambiguous Install Lifecycle Script

Install-time lifecycle script is not statically allowlisted and needs review.

package.jsonView on unpkg

mcp/launch.mjsView file

High

Known Malware Source Similarity

Source file is highly similar to a previously finalized malicious package; route for source-aware review.

mcp/launch.mjsView on unpkg

Findings

1 Critical4 High6 Medium7 Low

CriticalAi Agent Control Hijackmcp/install.mjs

HighInstall Time Lifecycle Scriptspackage.json

HighPayload In Excluded Dirgpl/fixtures/make_world_fixture.py

HighSemantic Analysis Limitedmcp/blender-server.mjs#virtual:normalized:round1

HighKnown Malware Source Similaritymcp/launch.mjs

MediumAmbiguous Install Lifecycle Scriptpackage.json

MediumDynamic Requiremcp/launch.mjs

MediumNetwork

MediumEnvironment Vars

MediumShips Build Helpergpl/convert.py

MediumStructural Risk Force Deep Review

LowScripts Present

LowWeak Cryptoscripts/apply-three-patch.mjs

Affected versions and remediation

This report applies to @threenative/core@0.3.6.

See all versions of @threenative/core for other recorded verdicts.

Review the evidence and your use of @threenative/core@0.3.6 before allowing it. Restrict the permissions described in this report.
Choose an independently verified alternative or release. This report does not establish that other versions are safe.

Evidence last updated: 2026-10-21 10:18:00.295Z (UTC).

Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Runtime surface

Behavioral surface

Source & flagged code