Security audit
Security checks across malware telemetry and agentic risk
This iMessage auto-replier mostly matches its stated purpose, but it needs review because it can read private texts, send them to AI services, text people automatically, reuse other tool API keys, and exposes an unsafe Telegram command path.
Install only if you are comfortable granting Full Disk Access and Messages automation to a tool that can read message history, send it to an AI provider, and send replies on your behalf. Use a dedicated API key for this skill, restrict the watch list, set nonzero delays, daily caps, time windows, and narrow keywords, and avoid untrusted Telegram/natural-language management input until the execSync command construction is fixed. Do not use it for sensitive conversations unless you accept the plaintext logging and off-device AI processing.
return new Promise((resolve, reject) => {
const proc = spawn('curl', [
'-s',
'https://api.openai.com/v1/chat/completions',
'-H', 'Content-Type: application/json',
'-H', `Authorization: Bearer ${apiKey}`,
'-d', payloadreturn new Promise((resolve, reject) => {
const proc = spawn('curl', [
'-s',
'https://api.anthropic.com/v1/messages',
'-H', 'Content-Type: application/json',
'-H', `x-api-key: ${apiKey}`,
'-H', 'anthropic-version: 2023-06-01',return new Promise((resolve, reject) => {
const proc = spawn('curl', [
'-s',
'https://api.openai.com/v1/chat/completions',
'-H', 'Content-Type: application/json',
'-H', `Authorization: Bearer ${apiKey}`,
'-d', payload66/66 vendors flagged this skill as clean.
Detected: suspicious.dangerous_exec, suspicious.exposed_secret_literal