@whalent/agent-core@0.3.466
Core runtime for Whalent Agent
AI Security Review
scanned 25d ago · by lpm-firewall-aiThe package can write foreign agent skill directories, but only after an explicit skills tool call, not from an npm lifecycle hook. That is a guarded agent-configuration capability, so it warrants a warning rather than a publish block.
Static reason
One or more suspicious static signals were detected.
Trigger
An explicit skills tool invocation on the running whalent-core process, with a platform and slug.
Impact
A consented agent session can add or replace skill content in Claude or Cursor control directories. No install-time mutation or proven silent exfiltration was found.
Mechanism
handleTool downloads files using a caller-supplied base and bearer token, then installSkill resolves a machine or workspace skill directory and writes the files.
AI rationale
The package can write foreign agent skill directories, but only after an explicit skills tool call, not from an npm lifecycle hook. That is a guarded agent-configuration capability, so it warrants a warning rather than a publish block.
Evidence
dist/mcp.cjsdist/index.cjs.claude.cursor
Decision evidence
public snapshotAI called this Suspicious at 74.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- package.json exposes bin whalent-core to dist/index.cjs and has no install lifecycle scripts.
- dist/mcp.cjs platformDir maps Claude platforms to .claude and Cursor platforms to .cursor.
- A tool handler downloads skill files with an Authorization header and calls installSkill using the caller platform.
- installSkill creates the resolved skill directory recursively and writes skill files there.
Evidence against
- Dependencies are ssh2 plus optional node-pty, not a self-referential release chain.
- No preinstall, install, or postinstall hook runs this write during npm install.
- Skill installation is inside an explicit tool switch, not module top-level execution.
Behavioral surface
CryptoDynamicRequireEnvironmentVarsEvalFilesystemNetwork
HighEntropyStringsMinifiedObfuscatedUrlStrings
Source & flagged code
7 flagged · loading sourcedist/mcp.cjsView file
1#!/usr/bin/env node
L2: 'use strict';var a0_0xc0d6b3=a0_0x256f;(function(_0x52673a,_0x6f0968){var _0x2c11cd=a0_0x256f,_0x3e63f0=_0x52673a();while(!![]){try{var _0x4d05f3=parseInt(_0x2c11cd(0xfcb))/0x1+par...
L3: || (${_0x4dbc28} == "string" && ${_0x5d353b} && ${_0x5d353b} == +${_0x5d353b})`)['assign'](_0xc7237d,(0x0,_0x5bc511['_'])`+${_0x5d353b}`);return;case'integer':_0x55db35[_0x4bc4b4(0...
High
Obfuscated Payload Loader
Source contains an obfuscated payload loader that reconstructs and executes hidden code.
dist/mcp.cjsView on unpkg · L1•stage = ast_semantic_analysis; reason = ast_path_work_budget_exceeded; limitedFiles = 1
High
Semantic Analysis Limited
A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.
dist/mcp.cjsView on unpkg4|| (${_0x4dbc28} === "string" && ${_0x5d353b} && ${_0x5d353b} == +${_0x5d353b} && !(${_0x5d353b} % 1))`)[_0x4bc4b4(0x1a09)](_0xc7237d,(0x0,_0x5bc511['_'])`+${_0x5d353b}`);return;ca...
L5: || ${_0x4dbc28} === "boolean" || ${_0x5d353b} === null`)['assign'](_0xc7237d,(0x0,_0x5bc511['_'])`[${_0x5d353b}]`);}}}function _0x5f47d3({gen:_0x175973,parentData:_0x43f5df,parentD...
L6: missingProperty: ${_0x48cec0},
Medium
Findings
3 High7 Medium5 Low
HighObfuscated Payload Loaderdist/mcp.cjs
HighObfuscated
HighSemantic Analysis Limiteddist/mcp.cjs
MediumDynamic Requiredist/mcp.cjs
MediumNetwork
MediumEnvironment Vars
MediumShips Build Helperdist/jupyter_sidecar.py
MediumOversized Source Filedist/index.cjs
MediumOversized Cli Entrypointdist/index.cjs
MediumStructural Risk Force Deep Review
LowEvaldist/mcp.cjs
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowNo License
Affected versions and remediation
This report applies to @whalent/agent-core@0.3.466.
See all versions of @whalent/agent-core for other recorded verdicts.
- Review the evidence and your use of @whalent/agent-core@0.3.466 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.