Shell command execution detected (child_process).
- Code
- suspicious.dangerous_exec
- Location
- dist/chunk-RWZ5QS2I.js:26
- Evidence
const r = spawnSync(cmd, args, { stdio: "inherit" });
Security audit
Security checks across malware telemetry and agentic risk
This is a powerful, mostly coherent memory plugin, but it needs Review because some optional tools can access or retain private project and prompt data more broadly than their docs indicate.
Install only if you want a heavy, persistent cognition stack and can operate it like infrastructure. Keep exposedTools on curated unless you have reviewed the all-tools surface, use a dedicated single-tenant database where possible, avoid logging sensitive prompts through ethics_audit until the retention mismatch is fixed, and verify Docker/network endpoints and credentials before production use.
62/62 vendors flagged this plugin as clean.
Detected: suspicious.dangerous_exec, suspicious.dynamic_code_execution, suspicious.env_credential_access (+3 more)
const r = spawnSync(cmd, args, { stdio: "inherit" });const makeValidate = new Function(`${names_1.default.self}`, `${names_1.default.scope}`, sourceCode);var ATLAS_URL = (process.env["CELIUMS_ATLAS_URL"] ?? "https://atlas.celiums.ai").replace(/\/$/, "");
const circadianOff = process.env["CELIUMS_RECALL_DISABLE_CIRCADIAN"] === "1";
var SEARCH_URL = (process.env["CELIUMS_SEARCH_URL"] ?? "").replace(/\/$/, "");
this.baseUrl = (cfg.baseUrl ?? process.env.INFERENCE_URL ?? "https://inference.do-ai.run/v1").replace(/\/+$/, "");
function llmConfigured(env = process.env) {...process.env,
client_secret: [REDACTED],
password: [REDACTED],
ssl: /sslmode=require/i.test(cs) || /ondigitalocean\.com/i.test(cs) ? { rejectUnauthorized: false } : void 0ssl: config.postgres.ssl ? { rejectUnauthorized: false } : void 0test: ["CMD-SHELL", "curl -sf http://127.0.0.1:9200/_cluster/health || exit 1"]