@tekmidian/pai@0.36.2
PAI Knowledge OS — Personal AI Infrastructure with federated memory and project management
AI Security Review
scanned 11d ago · by lpm-firewall-aiSource review found an interactive, consent-gated agent configuration capability and no install-time trigger or concrete attack behavior. This warrants a warning under the agent capability policy, not a malicious verdict.
Static reason
One or more suspicious static signals were detected.
Trigger
The user runs the package setup command and confirms the settings patch.
Impact
The configuration enables PAI hooks in later Claude sessions; the inspected flow asks before making this change.
Mechanism
The program merges missing hook registrations and other settings into ~/.claude/settings.json.
AI rationale
Source review found an interactive, consent-gated agent configuration capability and no install-time trigger or concrete attack behavior. This warrants a warning under the agent capability policy, not a malicious verdict.
Evidence
dist/program-C-fUghPv.mjs~/.claude/settings.json
Decision evidence
public snapshotAI called this Suspicious at 88.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The flagged hook files contain package functionality for reading project and session data, storing local notes, or communicating with the local daemon; their inspected target locations do not establish an attack.
- package.json defines only a prepublishOnly lifecycle script, which runs the build.
- The setup command asks before patching Claude settings with PAI hooks and other settings.
- After setup consent, the program merges hook registrations into Claude settings when values are missing.
Evidence against
- No consumer-install lifecycle hook was found in package.json.
- The setup prompt permits skipping the settings patch; the inspected source does not establish unconsented control-surface mutation.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetworkShell
HighEntropyStrings
Source & flagged code
10 flagged · loading sourcedist/context-handover-cache-PtNvj_8D.mjsView file
636}
L637: const { spawn } = await import("node:child_process");
L638: return new Promise((resolve) => {
High
Child Process
Package source references child process execution.
dist/context-handover-cache-PtNvj_8D.mjsView on unpkg · L636dist/program-C-fUghPv.mjsView file
3786],
L3787: shell: true
L3788: });
High
•Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation.
l.template.md");
if (!existssync(templatepath)) {
console.log(c.warn("skill template not found: " + templatepath));
console.log(c.dim(" skipping pai skill installation."));
return false;
}
const skilldir = join(homedir(), ".claude", "skills", "pai");
const skillfile = join(skilldir, "skill.md");
if (existssync(skillfile)) {
const content = readfilesync(skillfile, "utf-8");
const isgenerated = content.includes("generated by pai setup");
if (isgenerated) console.log(c.dim(" found existing pai-generated skill.md."));
else {
console.log(c.yellow(" found existing skill.md (not pai-generated)."));
console.log(c.dim(" a backup will be created before overwri
Findings
6 High6 Medium6 Low
HighChild Processdist/context-handover-cache-PtNvj_8D.mjs
HighShelldist/program-C-fUghPv.mjs
HighExternal Ai Agent Control Surface Mutationdist/hooks/capture-session-summary.mjs
HighTrigger Reachable External Ai Agent Control Surface Mutationdist/program-C-fUghPv.mjs
HighCross File Remote Execution Contextdist/program-C-fUghPv.mjs
HighTrigger Reachable Persistencedist/program-C-fUghPv.mjs
MediumDynamic Requirescripts/build-docs.mjs
MediumNetwork
MediumEnvironment Vars
MediumInstall Persistencedist/program-C-fUghPv.mjs
MediumShips Build Helpertab-color-command.sh
MediumStructural Risk Force Deep Review
Affected versions and remediation
This report applies to @tekmidian/pai@0.36.2.
See all versions of @tekmidian/pai for other recorded verdicts.
- Review the evidence and your use of @tekmidian/pai@0.36.2 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.