Is Mintplex-Labs/anything-llm safe?
- Possible command injection (exec with dynamic command)
- Dependency fetched from a git repository
- Node.js shell/command execution
What to do: Nothing here argues against installing it. Grant the capabilities it lists only if you expect the tool to need them.
Mintplex-Labs/anything-llm is a GitHub repository analyzed by SkillTotal's deterministic static scanner. The scan found no malicious indicators, though 2 risky constructs are reported for review. It can: delegated authentication, filesystem read, filesystem write, install time execution, mcp tools detected, network egress and shell execution — capabilities are what the code can do, not a verdict on intent. Risk score 23/100 (low).
anything-llm 1.19.0
Get notified when this component changes
Change alerts aren't live yet. Leave your email and we'll let you know when they launch. We use it only for this and delete it on request.
Automated static-analysis result. It can contain false positives and false negatives, and is not a claim about the intent of Mintplex-Labs/anything-llm's authors. Report a false positive.
Behavioral traits
How this component maps to the CSA agentic threat model. Descriptive — it never affects the risk score.
Findings (8)
The code builds an OS command out of values that can change at runtime, then runs it through a shell.
const result = execSync(`${which} ffmpeg`, { encoding: "utf8" }).trim();execSync(`"${pathToTest}" -version`, { encoding: "utf8", stdio: "pipe" });Why it matters: If any of those values come from untrusted input, an attacker can run their own commands on the machine.
Fix: Use execFile/spawn with an argument array instead of exec; never build a shell command string from external input.
The component can run operating-system commands or spawn processes.
const { execSync, spawnSync } = require("child_process");const result = execSync(`${which} ffmpeg`, { encoding: "utf8" }).trim();execSync(`"${pathToTest}" -version`, { encoding: "utf8", stdio: "pipe" });const result = spawnSync(
return require("child_process")const { spawn, execFile } = require("child_process");this.#spawn();
static #spawn() {const child = spawn(
const { spawnSync } = require("child_process");const result = spawnSync(rgPath, args, {const { spawnSync } = require("child_process");const result = spawnSync(rgPath, args, {const { spawn } = require("child_process");child = spawn(rgPath, args, { windowsHide: true });Why it matters: Powerful and often legitimate — confirm the commands aren't built from untrusted input.
Fix: Confirm the command and its arguments are fully controlled and not derived from untrusted input; prefer execFile with an argument array.
The component reads files from disk.
const mails = await mboxParser(fs.createReadStream(fullFilePath))
const buffer = await fs.readFile(this.filePath);
content = decodeText(fs.readFileSync(fullFilePath));
let buffer = fs.readFileSync(filePath);
file: fs.createReadStream(fullFilePath),
file: fs.createReadStream(fullFilePath),
buffer = fs.readFileSync(outputFile);
return fs.readFileSync(path.resolve(this.#storageLoc, this.#pubKeyName));
const content = fs.readFileSync(file, "utf-8");
const content = fs.readFileSync(sourceFile, "utf-8");
fs.readFileSync(path.resolve(__dirname, "index.css")),
fs.readFileSync(path.resolve(__dirname, "dark-swagger.css")),
fs.readFileSync(path.resolve(__dirname, "index.js"), "utf8"),
fs.readFileSync(path.resolve(__dirname, "index.js"), "utf8"),
const timestampMs = Number(fs.readFileSync(this.cacheAtPath));
fs.readFileSync(this.cacheModelPath, { encoding: "utf-8" }),fs.readFileSync(cacheModelPath, { encoding: "utf-8" }),const timestampMs = Number(fs.readFileSync(this.cacheAtPath));
fs.readFileSync(this.cacheModelPath, { encoding: "utf-8" }),fs.readFileSync(cacheModelPath, { encoding: "utf-8" }),fs.readFileSync(marker, "utf8").trim() === LEMOND_VERSION
config = JSON.parse(fs.readFileSync(configPath, "utf8"));
const versions = JSON.parse(fs.readFileSync(file, "utf8"));
const raw = fs.readFileSync("/sys/fs/cgroup/memory.max", "utf8").trim();const timestampMs = Number(fs.readFileSync(this.cacheAtPath));
Why it matters: Usually legitimate, but worth confirming it can't be steered into reading sensitive files.
Fix: Confirm which files are read and that paths cannot be influenced by untrusted input to reach sensitive locations.
The component writes or deletes files on disk.
fs.rmSync(outputFile);
const writeStream = fs.createWriteStream(localFilePath);
fs.rmSync(filepath);
fs.writeFileSync(destinationFilePath, JSON.stringify(data, null, 4), {fs.rmSync(path.join(directory, file));
fs.rmSync(path.join(directory, file));
fs.writeFileSync(
fs.writeFileSync(filePath, output, "utf-8");
fs.writeFileSync(
if (fs.existsSync(oldPfpPath)) fs.unlinkSync(oldPfpPath);
if (fs.existsSync(oldPfpPath)) fs.unlinkSync(oldPfpPath);
if (isDirectory) fs.rmSync(itemPath, { recursive: true });else fs.unlinkSync(itemPath);
fs.unlinkSync(path.join(generatedImagesPath, name));
fs.rmSync(filePath, { recursive: true });else fs.unlinkSync(filePath);
fs.unlinkSync(filePath);
fs.writeFileSync(destinationFilePath, JSON.stringify(jsonContent, null, 4), {fs.unlinkSync(sourceFile);
fs.writeFileSync(outputFile, JSON.stringify(openApiSpec, null, 2), {fs.writeFileSync(
fs.writeFileSync(
fs.writeFileSync(
fs.writeFileSync(
fs.rmSync(staging, { recursive: true, force: true });Why it matters: Usually legitimate, but worth confirming the paths can't be controlled by untrusted input.
Fix: Confirm which files are written/deleted and that paths cannot be influenced by untrusted input.
The component makes outbound network requests.
const pageText = await fetch(link, {const res = await fetch(url, {const res = await fetch(url, { signal: abortController.signal })const response = await fetch(url, { headers: initialHeaders });const response = await fetch(url, {const attachmentResponse = await fetch(downloadUrl, {const data = await fetch(nextUrl, {const response = await fetch(
const valid = await fetch(`${this.apiBase}/api/v1/user`, {const response = await fetch(url, {const response = await fetch(`${this.apiBase}/api/v1${endpoint}`, {const valid = await fetch(`${this.apiBase}/octocat`, {return fetch(`${this.apiBase}/repos/${this.author}/${this.project}`, {await fetch(`${this.apiBase}/api/v4/user`, {return fetch(`${this.apiBase}/api/v4/projects/${this.projectId}`, {const response = await fetch(url, {const response = await fetch(url, {const response = await fetch(url, {const videoResponse = await fetch(
const response = await fetch(
const response = await fetch(`http://127.0.0.1:11434/api/chat`, {return await fetch(`${API_BASE}/utils/embedded-lemonade/status`, {return await fetch(`${API_BASE}/utils/embedded-lemonade/recommendations`, {Why it matters: Usually legitimate, but confirm the destinations are expected and no sensitive data leaves.
Fix: Confirm the destination hosts are expected and that no sensitive data is sent off-host.
An OAuth 2.0 / OpenID Connect delegated-authentication flow was detected (authorization-code / refresh-token / token-exchange grant, an OIDC authorize/discovery endpoint or id_token, or a delegation library). Tools authenticate with the end user's delegated, scoped credentials rather than a long-lived embedded service credential. (2 occurrence(s) shown as evidence).
grant_type: "authorization_code",
grant_type: "refresh_token",
Fix: Delegated auth is a lower-blast-radius execution context than an embedded static credential. Confirm the requested scopes are minimal and that tokens are never logged or forwarded off-host.
An MCP tool surface (manifest or tool definitions) was found.
Why it matters: Just context — review which tools it offers and their permissions.
Fix: Review the declared MCP tools and their permissions.
package.json lists a dependency as a git or GitHub source. npm clones it at install time: the code is not a registry release and is not part of this scan.
"epub2": "git+https://github.com/Mintplex-Labs/epub2-static.git#main",
Fix: Prefer a registry release; otherwise pin the dependency to a commit hash.
How attackers abuse these capabilities
Interactive labs on the attack class behind the rules above. They show the technique, not anything found in Mintplex-Labs/anything-llm.
Check your own component
Run the same evidence-backed scan on any MCP server, agent skill, or package.
Scan your own componentHow we determine this: deterministic static analysis (regex + AST), evidence-anchored, no code execution. Methodology →