memgineering@0.22.0
One memory for the AI you connect. Recall, remember, and revise a brain your agents share — stored in your own folder.
AI Security Review
scanned 2mo ago · by lpm-firewall-aiThe package does not directly mutate agent configuration during postinstall, but it uses an automatic lifecycle hook to pressure an AI agent into initiating a command that can make broad global control-surface changes. The alleged credential redirect was not confirmed; the observed default endpoint is package-aligned.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
Global npm installation followed by an agent obeying the postinstall output.
Impact
Can induce configuration changes in Claude or Codex control surfaces without a direct user command.
Mechanism
Lifecycle-delivered agent instruction leading to global agent configuration writes.
AI rationale
The package does not directly mutate agent configuration during postinstall, but it uses an automatic lifecycle hook to pressure an AI agent into initiating a command that can make broad global control-surface changes. The alleged credential redirect was not confirmed; the observed default endpoint is package-aligned.
Evidence
package.jsonscripts/postinstall.mjsdist/index.js~/.claude/CLAUDE.md~/.claude/skills~/.claude/settings.json~/.codex/AGENTS.md~/.codex/skills~/.codex/hooks.json
Network endpoints2
api.memgineering.commemgineering.com
Decision evidence
public snapshotAI called this Suspicious at 91.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- A postinstall hook runs automatically for global installs and instructs an AI agent to run setup immediately.
- The setup command writes guidance and skills into global Claude and Codex control files.
- The package sends setup and authenticated API requests to its hosted API.
Evidence against
- The postinstall script itself only prints when npm marks the install as global; it does not directly run setup.
- No self-dependency or fixed third-party credential gateway was found.
- The default API host matches the package domain.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell
HighEntropyStringsUrlStrings
Source & flagged code
7 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node scripts/postinstall.mjs
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node scripts/postinstall.mjs
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgdist/index.jsView file
15function brandHome() {
L16: return process.env[BRAND.homeEnvVar] ?? join(homedir(), BRAND.homeDirName);
L17: }
...
L121: function isInteractive() {
L122: return Boolean(process.stdout.isTTY) && !jsonMode;
L123: }
L124: function printHuman(text) {
L125: if (!jsonMode) process.stdout.write(text + "\n");
L126: }
...
L137: for (let i = 0; i < input.length; i++) {
L138: const code = input.charCodeAt(i);
L139: if (code === 27 && input[i + 1] === "]" || code === 157) {
Critical
Credential Exfiltration
Findings
3 Critical3 High4 Medium5 Low
CriticalCredential Exfiltrationdist/index.js
CriticalTrigger Reachable Dangerous Capabilitydist/index.js
CriticalPrevious Version Dangerous Deltadist/index.js
HighInstall Time Lifecycle Scriptspackage.json
HighCredential Redirect Persistencedist/index.js
HighTrigger Reachable Credential Exfiltrationdist/index.js
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
LowNon Install Lifecycle Scripts
LowScripts Present
LowFilesystem
Low
Affected versions and remediation
This report applies to memgineering@0.22.0.
See all versions of memgineering for other recorded verdicts.
- Review the evidence and your use of memgineering@0.22.0 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.