@tekmidian/pai@0.42.1
PAI Knowledge OS — Personal AI Infrastructure with federated memory and project management
AI Security Review
scanned 11d ago · by lpm-firewall-aiThe package has no automatic install hook and the inspected Claude settings mutation is performed through interactive, consent-guarded setup. This first-party agent configuration capability warrants a warning, but the inspected source does not establish malicious behavior.
Static reason
One or more suspicious static signals were detected.
Trigger
A user runs the setup command and accepts the settings patch prompt.
Impact
This changes AI-agent behavior for the user's Claude installation after consent; the inspected source does not establish covert control or credential theft.
Mechanism
The package merges PAI hooks, environment values, permissions, and flags into Claude settings, adding missing values without replacing existing ones.
AI rationale
The package has no automatic install hook and the inspected Claude settings mutation is performed through interactive, consent-guarded setup. This first-party agent configuration capability warrants a warning, but the inspected source does not establish malicious behavior.
Evidence
dist/program-B7N5OIJR.mjs~/.claude/settings.json
Decision evidence
public snapshotAI called this Suspicious at 86.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The manifest has no install-time lifecycle hook; its listed prepublishOnly command runs the build.
- The inspected hooks read configuration from PAI or Claude directories, and the context loader locates project notes under Claude project paths.
- Interactive setup offers to add PAI hooks, environment values, permissions, and flags to Claude settings, then asks before applying the change.
- The worker provider command takes its endpoint and key-file path from explicit command options.
Evidence against
- The settings patch is guarded by an interactive consent prompt, and the merge routine adds missing values without overwriting existing ones.
- No source evidence inspected establishes unconsented credential forwarding or an active attack.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetworkShell
Source & flagged code
11 flagged · loading sourcedist/router-DD6D4j31.mjsView file
2import { connect } from "node:net";
L3: import { spawn } from "node:child_process";
L4:
High
Child Process
Package source references child process execution.
dist/router-DD6D4j31.mjsView on unpkg · L2dist/program-B7N5OIJR.mjsView file
3803],
L3804: shell: true
L3805: });
High
•Manifest-reachable source captures an API credential, sends it to a fixed unofficial gateway, and persists that redirection in agent or shell configuration.
dist/program-B7N5OIJR.mjs:
import { appendFileSync, chmodSync, copyFileSync, createReadStream, existsSync, lstatSync, mkdirSync, readFileSync, readdirSync, readlinkSync, realpathSync, renameSync, statSync, s...
import { createInterface } from "node:readline";
* 08 - Others/MDF/Infrastruktur/.claude/worktrees/cool-haibt 1 session
* 08 - Others/MDF/Infrastruktur/.claude/worktrees/strange-haibt 7 sessions
fragment: `${sep}.claude${sep}worktrees${sep}`,
writeFileSync(join(targetPath, "Notes", "MEMORY.md"), memoryContent, "utf-8");
const claudeProjects = join(homedir(), ".claude", "projects");
writeFileSync(configPath, JSON.strin
Findings
7 High6 Medium6 Low
HighChild Processdist/router-DD6D4j31.mjs
HighShelldist/program-B7N5OIJR.mjs
HighCredential Redirect Persistencedist/program-B7N5OIJR.mjs
HighExternal Ai Agent Control Surface Mutationdist/hooks/capture-session-summary.mjs
HighTrigger Reachable External Ai Agent Control Surface Mutationdist/program-B7N5OIJR.mjs
HighCross File Remote Execution Contextdist/program-B7N5OIJR.mjs
HighTrigger Reachable Persistencedist/program-B7N5OIJR.mjs
MediumDynamic Requirescripts/build-docs.mjs
MediumNetwork
MediumEnvironment Vars
MediumInstall Persistencedist/program-B7N5OIJR.mjs
MediumShips Build Helper
Affected versions and remediation
This report applies to @tekmidian/pai@0.42.1.
See all versions of @tekmidian/pai for other recorded verdicts.
- Review the evidence and your use of @tekmidian/pai@0.42.1 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.