@whalent/agent-core@0.3.479
Core runtime for Whalent Agent
AI Security Review
scanned 23d ago · by lpm-firewall-aiSource establishes tool-invoked agent control-surface mutation, which the supplied policy treats as a warning. No block-eligible malicious chain was established; heavy entrypoint obfuscation limits broader assurance.
Static reason
One or more suspicious static signals were detected.
Trigger
An MCP client invokes the skill installation or MCP registration tool.
Impact
Invoked tools can change agent instructions and registered executable commands.
Mechanism
Tool arguments select skill scope or MCP command settings; handlers write skill files or update agent configuration.
AI rationale
Source establishes tool-invoked agent control-surface mutation, which the supplied policy treats as a warning. No block-eligible malicious chain was established; heavy entrypoint obfuscation limits broader assurance.
Evidence
dist/mcp.cjsconfig.toml.claude.jsonopencode.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 exposes a CLI and contains no install lifecycle hooks or runtime self-dependency.
- The MCP request handler dispatches caller-selected tools and arguments.
- The MCP management tool accepts command, arguments, and environment settings and registers them in agent configuration.
- The Codex configuration writer updates config.toml.
- The skill management tool installs downloaded files into selectable scopes, including the user's platform skills directory.
Evidence against
- The inspected configuration mutations are tool-invoked rather than npm-install-triggered.
- No concrete covert credential export or malicious payload execution was established.
Behavioral surface
CryptoDynamicRequireEnvironmentVarsEvalFilesystemNetwork
HighEntropyStringsMinifiedObfuscatedUrlStrings
Source & flagged code
7 flagged · loading sourcedist/mcp.cjsView file
1#!/usr/bin/env node
L2: 'use strict';var a0_0x237aec=a0_0x56ff;(function(_0x4ea4b8,_0x3d7b92){var _0xe3f083=a0_0x56ff,_0x53d9c6=_0x4ea4b8();while(!![]){try{var _0x9399c2=parseInt(_0xe3f083(0x1755))/0x1+pa...
L3: || (${_0x3ac3e8} == "string" && ${_0x1b5e3e} && ${_0x1b5e3e} == +${_0x1b5e3e})`)['assign'](_0x14f84b,(0x0,_0x1a8a18['_'])`+${_0x1b5e3e}`);return;case _0x4cbbe9(0x22b4):_0x1a489f[_0...
High
Obfuscated Payload Loader
Source contains an obfuscated payload loader that reconstructs and executes hidden code.
dist/mcp.cjsView on unpkg · L1•stage = ast_semantic_analysis; reason = ast_path_work_budget_exceeded; limitedFiles = 1
High
Semantic Analysis Limited
A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.
dist/mcp.cjsView on unpkg4|| (${_0x3ac3e8} === "string" && ${_0x1b5e3e} && ${_0x1b5e3e} == +${_0x1b5e3e} && !(${_0x1b5e3e} % 1))`)['assign'](_0x14f84b,(0x0,_0x1a8a18['_'])`+${_0x1b5e3e}`);return;case'boolea...
L5: || ${_0x3ac3e8} === "boolean" || ${_0x1b5e3e} === null`)[_0x4cbbe9(0x573)](_0x14f84b,(0x0,_0x1a8a18['_'])`[${_0x1b5e3e}]`);}}}function _0xd02005({gen:_0x372e8c,parentData:_0x5cf7fa...
L6: missingProperty: ${_0xfe9d5},
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@0.3.479.
See all versions of @whalent/agent-core for other recorded verdicts.
- Review the evidence and your use of @whalent/agent-core@0.3.479 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.