codex-supermemory@1.0.19
Persistent memory for OpenAI Codex CLI — powered by Supermemory
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe flagged credential requests appear consistent with the package’s Supermemory API use. However, automatic authentication sends the hostname and working-directory path to the vendor without disclosing those fields, so this merits a telemetry warning rather than a malicious verdict.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
A Codex session starts while no Supermemory API key is configured.
Impact
Supermemory receives the machine hostname and local working-directory path during authentication.
Mechanism
The hook automatically launches the Supermemory auth URL with the machine hostname, OS details, and current working directory as query parameters.
AI rationale
The flagged credential requests appear consistent with the package’s Supermemory API use. However, automatic authentication sends the hostname and working-directory path to the vendor without disclosing those fields, so this merits a telemetry warning rather than a malicious verdict.
Evidence
dist/hooks/session-start.jsREADME.md~/.codex/supermemory/credentials.json
Network endpoints2
console.supermemory.ai/auth/connectapi.supermemory.ai
Decision evidence
public snapshotAI called this Suspicious at 83.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- When no API key is configured, the session-start hook automatically opens the authentication flow and includes the machine hostname and current working directory in the auth URL.
- The README discloses that browser authentication opens on the first prompt, but does not mention sending the hostname or working directory.
Evidence against
- The auth flow targets Supermemory’s own authentication service and is part of the documented login process.
- The flagged API request sends the Supermemory key to the configured Supermemory API base URL for container-tag operations.
- The package has no install lifecycle scripts.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetwork
HighEntropyStringsUrlStrings
Source & flagged code
6 flaggeddist/hooks/recall.jsView file
177var import_node_crypto = require("node:crypto");
L178: var import_node_child_process = require("node:child_process");
L179: var import_node_fs3 = require("node:fs");
High
Child Process
Package source references child process execution.
dist/hooks/recall.jsView on unpkg · L177dist/cli.jsView file
413try {
L414: const utilInspect = eval("require('util').inspect");
L415: _custom = utilInspect.custom;
High
dist/hooks/mcp-proxy.jsView file
4// src/hooks/mcp-proxy.ts
L5: var import_node_readline = require("node:readline");
L6:
Findings
1 Critical4 High4 Medium4 Low
CriticalCredential Exfiltrationdist/hooks/session-start.js
HighChild Processdist/hooks/recall.js
HighEvaldist/cli.js
HighSame File Env Network Executiondist/hooks/session-start.js
HighSandbox Evasion Gated Capabilitydist/hooks/session-start.js
MediumDynamic Requiredist/hooks/mcp-proxy.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowScripts Present
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to codex-supermemory@1.0.19.
See all versions of codex-supermemory for other recorded verdicts.
- Review the evidence and your use of codex-supermemory@1.0.19 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: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.