registry / engine7 / 7.1.68

engine7@7.1.68

Engine 7 — 给助手一个家 / A home for your AI assistant

AI Security Review

scanned 1mo ago · by lpm-firewall-ai

No confirmed automatic attack surface was established. Inspected external requests are runtime agent features rather than installation behavior.

Static reason

High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review; previous stored version diff introduced dangerous source

Trigger

Explicitly starting or using the agent's configured runtime features.

Impact

No unsolicited credential collection, exfiltration, or install-time mutation was confirmed.

Mechanism

User-invoked agent service calls.

AI rationale

The package has no registry-install lifecycle hook, and the flagged code is consistent with an explicitly run AI-agent application and bundled utilities. Its broad runtime capabilities warrant normal operational caution, but static inspection found no concrete malicious delivery or exfiltration chain.

Decision evidence

public snapshot

AI called this Clean at 88.0% confidence as Benign with low false-positive risk.

Why the final policy warns

The AI recommended clean, but the final policy action is warn. This report does not record a matching static-policy override reason.

Evidence for AI clean decision

The runtime image-edit feature posts its generated request body to an external image API using a configured key.
The main entry point starts the engine only after loading command-line configuration.
The init template can create a launcher that resolves a local or global engine7 installation.
The bundled brainstorming server defaults to a loopback bind and contains a static branding image URL.
The bundled graph renderer invokes Graphviz on diagram content supplied through its command-line workflow.
The manifest's only lifecycle script is prepublishOnly, not an install-time hook.

Evidence against

No preinstall, install, or postinstall script is present.
No inspected source path showed automatic credential harvesting or an unsolicited remote receiver.
The flagged template files are a local companion server and a user-run Graphviz renderer, not install-time payloads.

Behavioral surface

Source

ChildProcess

Source & flagged code

18 flagged

dist/engine-startup.mjsView file

265Source sends the broad process environment to a literal external destination. L265: try { L266: const raw = JSON.parse(fs2.readFileSync(filePath, "utf-8")); L267: loadHooksFromConfig(raw); ... L369: // src/hooks/executor.ts L370: import { spawn } from "child_process"; L371: import { randomUUID } from "node:crypto"; L372: function parseHookOutput(stdout) { L373: const trimmed = stdout.trim(); ... L450: const envVars = { L451: ...process.env, L452: CLAUDE_PROJECT_DIR: ctx.workspace ... L454: const command = hook.command;

Critical

Hardcoded Runtime Data Exfiltration

Source sends credentials or rich application records to a package-controlled external receiver enabled by default.

dist/engine-startup.mjsView on unpkg · L265

265try { L266: const raw = JSON.parse(fs2.readFileSync(filePath, "utf-8")); L267: loadHooksFromConfig(raw); ... L369: // src/hooks/executor.ts L370: import { spawn } from "child_process"; L371: import { randomUUID } from "node:crypto"; L372: function parseHookOutput(stdout) { L373: const trimmed = stdout.trim(); ... L450: const envVars = { L451: ...process.env, L452: CLAUDE_PROJECT_DIR: ctx.workspace ... L454: const command = hook.command;

Critical

Credential Exfiltration

Source appears to send environment or credential material to an external endpoint.

dist/engine-startup.mjsView on unpkg · L265

•matchType = previous_version_dangerous_delta matchedPackage = engine7@7.1.67 matchedIdentity = npm:ZW5naW5lNw:7.1.67 similarity = 0.917 summary = stored previous version shares package body but lacks this dangerous source file

Critical

Previous Version Dangerous Delta

This package version adds a dangerous source file absent from the previous stored version; route for source-aware review.

dist/engine-startup.mjsView on unpkg

369// src/hooks/executor.ts L370: import { spawn } from "child_process"; L371: import { randomUUID } from "node:crypto";

High

Child Process

Package source references child process execution.

dist/engine-startup.mjsView on unpkg · L369

3740filesystem: true, L3741: shell: true, L3742: memory: true,

High

Shell

Package source references shell execution.

dist/engine-startup.mjsView on unpkg · L3740

High

Command Output Exfiltration

Source combines command execution, command-output handling, and outbound requests; review data flow before blocking.

dist/engine-startup.mjsView on unpkg · L265

Low

Weak Crypto

Package source references weak cryptographic algorithms.

dist/engine-startup.mjsView on unpkg · L265

dist/cli.mjsView file

•Manifest-reachable source resolves another installed package, overwrites its runtime code, and injects package-defined remote behavior. dist/cli.mjs: feishu: "https://accounts.feishu.cn", lark: "https://accounts.larksuite.com" fs.writeFileSync(fp, JSON.stringify({ sync_buf: buf })); if (redirectHost) currentBaseUrl = `https://${redirectHost}`; fs.writeFileSync(credFile, JSON.stringify({ accountId, token, baseUrl, userId }, null, 2), "utf8"); ILINK_BASE_URL = "https://ilinkai.weixin.qq.com"; WEIXIN_CDN_BASE_URL = "https://novac2c.cdn.weixin.qq.com/c2c"; fs.writeFileSync(this.filePath, JSON.stringify(obj));

High

Entrypoint Foreign Package Code Overwrite

Manifest-reachable source overwrites another installed package with package-defined remote behavior.

dist/cli.mjsView on unpkg

48Cross-file remote execution chain: dist/cli.mjs spawns dist/main.mjs; helper contains network access plus dynamic code execution. L48: const url = `${ACCOUNTS_URL[domain]}${REGISTRATION_PATH}`; L49: const res = await fetch(url, { L50: method: "POST", L51: headers: { "Content-Type": "application/x-www-form-urlencoded" }, L52: body: new URLSearchParams(body).toString(), L53: signal: AbortSignal.timeout(REQUEST_TIMEOUT_MS) ... L55: if (!res.ok) { L56: throw new Error(`\u98DE\u4E66\u6CE8\u518C\u63A5\u53E3\u8FD4\u56DE ${res.status}: ${await res.text()}`); L57: } ... L201: import * as os from "os"; L202: import { promises as dns } from "dns"; L203: function sleep2(ms) {

High

Cross File Remote Execution Context

Source spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.

dist/cli.mjsView on unpkg · L48

48Trigger-reachable persistence chain: manifest.bin -> bin/engine7 -> dist/cli.mjs L48: const url = `${ACCOUNTS_URL[domain]}${REGISTRATION_PATH}`; L49: const res = await fetch(url, { L50: method: "POST", L51: headers: { "Content-Type": "application/x-www-form-urlencoded" }, L52: body: new URLSearchParams(body).toString(), L53: signal: AbortSignal.timeout(REQUEST_TIMEOUT_MS) ... L55: if (!res.ok) { L56: throw new Error(`\u98DE\u4E66\u6CE8\u518C\u63A5\u53E3\u8FD4\u56DE ${res.status}: ${await res.text()}`); L57: } ... L201: import * as os from "os"; L202: import { promises as dns } from "dns"; L203: function sleep2(ms) {

dist/main.mjsView file

99Trigger-reachable chain: scripts.start -> dist/main.mjs L99: try { L100: const raw = JSON.parse(fs51.readFileSync(cfgPath, "utf-8")); L101: const keys = dotPath.split("."); ... L136: filesystem: true, L137: shell: true, L138: memory: true, ... L496: import { randomUUID } from "node:crypto"; L497: function parseHookOutput(stdout) { L498: const trimmed = stdout.trim(); ... L575: const envVars = { L576: ...process.env, L577: CLAUDE_PROJECT_DIR: ctx.workspace

Critical

Trigger Reachable Dangerous Capability

A package entrypoint or install-time lifecycle script reaches a source file with blocking dangerous behavior.

dist/main.mjsView on unpkg · L99

•Manifest-reachable source captures an API credential, sends it to a fixed unofficial gateway, and persists that redirection in agent or shell configuration. dist/main.mjs: fs51.writeFileSync(cfgPath, JSON.stringify(raw, null, 2) + "\n", "utf-8"); fs4.writeFileSync(dumpPath, [systemStable, systemDynamic].join("\n\n"), "utf-8"); fs4.writeFileSync(path3.join(opts.workspace, ".context-debug.txt"), lines.join("\n") + "\n", "utf-8"); if (process.env.CLAUDE_CODE_MAX_RETRIES) { return parseInt(process.env.CLAUDE_CODE_MAX_RETRIES, 10); let base = (this.config.baseUrl || "").trim(); "Authorization": `Bearer ${this.config.apiKey}` let u = (config2.baseUrl || "https://api.anthropic.com").trim();

High

Credential Redirect Persistence

Manifest-reachable source sends a prompted API credential to a fixed unofficial gateway and persists the redirection.

dist/main.mjsView on unpkg

99Trigger-reachable credential exfiltration chain: scripts.start -> dist/main.mjs L99: try { L100: const raw = JSON.parse(fs51.readFileSync(cfgPath, "utf-8")); L101: const keys = dotPath.split("."); ... L136: filesystem: true, L137: shell: true, L138: memory: true, ... L496: import { randomUUID } from "node:crypto"; L497: function parseHookOutput(stdout) { L498: const trimmed = stdout.trim(); ... L575: const envVars = { L576: ...process.env, L577: CLAUDE_PROJECT_DIR: ctx.workspace

bin/engine7.cmdView file

•path = bin/engine7.cmd kind = build_helper sizeBytes = 42 magicHex = [redacted]

Medium

Ships Build Helper

Package ships non-JavaScript build or shell helper files.

bin/engine7.cmdView on unpkg

templates/skills/superpowers/brainstorming/scripts/server.cjs#virtual:normalized:round1View file

•stage = ast_semantic_analysis; reason = ast_parse_error; limitedFiles = 1

High

Semantic Analysis Limited

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

templates/skills/superpowers/brainstorming/scripts/server.cjs#virtual:normalized:round1View on unpkg

templates/skills/superpowers/brainstorming/scripts/server.cjsView file

•matchType = normalized_sha256 matchedPackage = @brainervirus/workit-opencode@0.8.5 matchedPath = [redacted].cjs matchedIdentity = npm:[redacted]:0.8.5 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.

templates/skills/superpowers/brainstorming/scripts/server.cjsView on unpkg

templates/skills/superpowers/writing-skills/render-graphs.jsView file

•matchType = normalized_sha256 matchedPackage = @abblor/agent-os@0.11.0 matchedPath = skills/writing-skills/render-graphs.js matchedIdentity = npm:QGFiYmxvci9hZ2VudC1vcw:0.11.0 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.

templates/skills/superpowers/writing-skills/render-graphs.jsView on unpkg

Source & flagged code

18 flagged · loading source

dist/engine-startup.mjsView file

Critical

Hardcoded Runtime Data Exfiltration

Source sends credentials or rich application records to a package-controlled external receiver enabled by default.

dist/engine-startup.mjsView on unpkg · L265

Critical

Credential Exfiltration

Source appears to send environment or credential material to an external endpoint.

dist/engine-startup.mjsView on unpkg · L265

Findings

4 Critical11 High5 Medium6 Low

CriticalHardcoded Runtime Data Exfiltrationdist/engine-startup.mjs

CriticalCredential Exfiltrationdist/engine-startup.mjs

CriticalTrigger Reachable Dangerous Capabilitydist/main.mjs

CriticalPrevious Version Dangerous Deltadist/engine-startup.mjs

HighChild Processdist/engine-startup.mjs

HighShelldist/engine-startup.mjs

HighCredential Redirect Persistencedist/main.mjs

HighEntrypoint Foreign Package Code Overwritedist/cli.mjs

HighCommand Output Exfiltrationdist/engine-startup.mjs

HighCross File Remote Execution Contextdist/cli.mjs

HighTrigger Reachable Credential Exfiltrationdist/main.mjs

Affected versions and remediation

This report applies to engine7@7.1.68.

See version security history for other recorded verdicts.

Review the evidence and your use of engine7@7.1.68 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: 2026-09-05 08:46:58.881Z (UTC).