Loading npm security reports…

@linghun/tui: Suspicious npm security report (Warn) | LPM Firewall

registry / @linghun/tui / 0.1.35

@linghun/tui@0.1.35

AI Security Review

scanned 2h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. The runtime can manage first-party `.linghun` extensions, start configured MCP processes, and deliver redacted remote notifications. These capabilities require runtime user/config activation; no install-time attack surface was found.

Static reason

No blocking static signals were detected.

Trigger

A user invokes `/skills`, `/plugins`, `/remote`, or configures/uses an MCP server in the TUI runtime.

Impact

A consenting user can add package-owned extensions, execute configured MCP servers, or send redacted summaries to configured messaging destinations.

Mechanism

Explicit agent-extension lifecycle, configured subprocess execution, and configured remote notification delivery.

Rationale

No lifecycle hook, covert exfiltration, hard-coded hostile endpoint, or unconsented foreign AI-agent control-surface mutation was found. Mark as warn for its guarded but high-impact first-party agent-extension and configured command-execution capabilities.

Evidence

package.jsondist/index.jsdist/chunk-3Y7WSQP5.jsdist/chunk-VAQEIQEQ.jsdist/chunk-BNLU7ME2.jsdist/chunk-KBQ6GH7G.js.linghun/settings.json

Decision evidence

public snapshot

AI called this Suspicious at 87.0% confidence as Dangerous Capability with low false-positive risk.

Evidence for warning

`dist/chunk-3Y7WSQP5.js` supports explicit `/skills` and `/plugins` install, enable, update, and removal commands.
`dist/chunk-VAQEIQEQ.js` can clone/fetch extension sources and install them into package-owned extension locations after `--confirm-network`.
`dist/chunk-VAQEIQEQ.js` persists extension enablement in `.linghun/settings.json`.
`dist/chunk-BNLU7ME2.js` sends configured redacted summaries to webhook endpoints or configured messaging CLIs.
`dist/chunk-KBQ6GH7G.js` spawns configured MCP stdio server commands at runtime.

Evidence against

`package.json` has no `preinstall`, `install`, `postinstall`, or other lifecycle hook.
`dist/index.js` exports runtime APIs; inspected entrypoint has no install-time invocation.
Extension network install requires an explicit slash command and `--confirm-network`.
Extension code is guarded by enabled/trusted/manifest/runtime checks in `dist/chunk-VAQEIQEQ.js`.
Webhook delivery sends `redactedSummary` only and uses user-configured endpoints.
No hard-coded non-example network endpoint, credential harvesting, stealth persistence, or destructive payload was found.

Source & flagged code

1 flagged

dist/chunk-6PZJEDSA.jsView file

Lines 246-286javascript

246 if (!reason) return void 0;

247 const parts = reason.split(";").filter(

248 (part) => part.startsWith("generatedEvidence=") || part === "generatedVerification=partial" || part === "index=unknown_nonblocking"

249 );

250 return parts.length > 0 ? parts.join(";") : void 0;

251}

252function deriveAgentDisplayName(type, task) {

253 const stopWords = /* @__PURE__ */ new Set([

254 "a",

255 "an",

256 "and",

257 "for",

258 "help",

259 "job",

260 "please",

261 "run",

262 "task",

263 "the",

264 "to",

265 "with"

266 ]);

267 const asciiTask = Array.from(task.normalize("NFKD").toLowerCase()).map((char) => char.charCodeAt(0) <= 127 ? char : " ").

268 const tokens = asciiTask.replace(/[^a-z0-9]+/gu, " ").trim().split(/\s+/u).filter((token) => token &&

269 const base = tokens.length > 0 ? tokens.join("-") : `task-${createHash("sha1").update(task).digest("hex").slice(

270 const label = base.endsWith(`-${type}`) ? base : `${base}-${type}`;

271 return truncateAsciiLabel(label, 36);

272}

273function truncateAsciiLabel(value, maxLength) {

274 const cleaned = value.replace(/[^a-z0-9-]/gu, "-").replace(/-{2,}/gu, "-").replace(/^-|-$/gu, "");

275 if (cleaned.length <= maxLength) {

276 return cleaned || "agent";

277 }

278 return cleaned.slice(0, maxLength).replace(/-+$/u, "") || "agent";

279}

280function createDurableJobAgents(options, status, runningCap) {

281 const total = Math.max(1, options.requestedAgents);

282 const tasks = createDurableJobAgentTasks(options.goal, total);

283 return Array.from({ length: total }, (_, index) => {

284 const active = false;

285 const agentStatus = status === "running" ? index < runningCap ? "queued" : "sleeping" : status === "created" ? "created" : status;

286 const type = index === 0 ? "planner" : index === 1 ? "worker" : index === 2 ? "verifier" :

Source & flagged code

1 flagged · loading source

dist/chunk-6PZJEDSA.jsView file

266]); L267: const asciiTask = Array.from(task.normalize("NFKD").toLowerCase()).map((char) => char.charCodeAt(0) <= 127 ? char : " ").join(""); L268: const tokens = asciiTask.replace(/[^a-z0-9]+/gu, " ").trim().split(/\s+/u).filter((token) => token && !stopWords.has(token)).slice(0, 3);

Low

Weak Crypto

Package source references weak cryptographic algorithms.

dist/chunk-6PZJEDSA.jsView on unpkg · L266

Findings

3 Medium6 Low

MediumNetwork

MediumEnvironment Vars

MediumStructural Risk Force Deep Review

LowScripts Present

LowWeak Cryptodist/chunk-6PZJEDSA.js

LowFilesystem

LowObfuscated

LowHigh Entropy Strings

LowUrl Strings