agenticros@0.7.33
AgenticROS - agentic AI for ROS-powered robots. Single CLI to launch real-robot or simulated demos, manage configuration, and inspect status.
AI Security Review
scanned 16d ago · by lpm-firewall-aiThe package has no npm install hook, but explicit setup commands mutate agent configuration and add an Antigravity allow rule. Under the agent integration policy, this warrants a warning for explicit-command agent configuration capability, not a block.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review
Trigger
A user runs an agent setup command such as `agenticros claude` or the Antigravity setup path.
Impact
The agent can launch the local AgenticROS MCP server and use its tools according to the agent configuration.
Mechanism
The setup writer adds the local MCP server registration; Antigravity setup also writes a skill and an allow rule in agent settings.
AI rationale
The package has no npm install hook, but explicit setup commands mutate agent configuration and add an Antigravity allow rule. Under the agent integration policy, this warrants a warning for explicit-command agent configuration capability, not a block.
Evidence
dist/util/mcp-setup.jsdist/commands/claude.jsdist/util/agy-config.js
Decision evidence
public snapshotAI called this Suspicious at 88.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The explicit Claude setup command invokes the shared MCP configuration writer.
- The Antigravity setup path writes a global MCP registration, may write a project registration, and adds an allow rule to agent settings.
- The MCP registration launches the local AgenticROS entrypoint with Node.
Evidence against
- The manifest defines only a prepublishOnly lifecycle script, not an install-time hook.
- The inspected agent configuration changes are reachable through setup commands, not an npm installation hook.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkShellWebSocket
HighEntropyStringsUrlStrings
Source & flagged code
24 flagged · loading sourcedist/util/mcp-discovery.jsView file
16*/
L17: import { spawn } from "node:child_process";
L18: import { existsSync } from "node:fs";
High
Child Process
Package source references child process execution.
dist/util/mcp-discovery.jsView on unpkg · L16•matchType = normalized_sha256
matchedPackage = agenticros@0.6.6
matchedPath = dist/util/mcp-discovery.js
matchedIdentity = npm:YWdlbnRpY3Jvcw:0.6.6
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High
Known Malware Source Similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
dist/util/mcp-discovery.jsView on unpkgdist/commands/init.js#virtual:normalized:round1View file
21import { confirm, input, password, select } from "@inquirer/prompts";
L22: import { execa } from "execa";
L23: import { runDoctorChecks } from "./doctor.js";
Findings
20 High6 Medium6 Low
HighChild Processdist/util/mcp-discovery.js
HighShelldist/commands/init.js#virtual:normalized:round1
HighSame File Env Network Executiondist/util/marketplace.js
HighTrigger Reachable External Ai Agent Control Surface Mutationdist/util/agy-config.js
HighRuntime Package Installdist/commands/init.js#virtual:normalized:round1
HighShips High Entropy Blobruntime/packages/agenticros/agenticros-agenticros-0.0.1.tgz
HighPayload In Excluded Dirruntime/ros2_ws/src/agenticros_explore/test/test_frontiers.py
HighSemantic Analysis Limiteddist/commands/init.js#virtual:normalized:round1
HighKnown Malware Source Similaritydist/commands/web.js
HighKnown Malware Source Similaritydist/util/mcp-discovery.js
Affected versions and remediation
This report applies to agenticros@0.7.33.
See all versions of agenticros for other recorded verdicts.
- Review the evidence and your use of agenticros@0.7.33 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.