ClawHub
Back to skill

Security audit

deploy-packager

Security checks across malware telemetry and agentic risk

Overview

This deployment packager needs review because it can read plan-selected files and publish deployment metadata while claiming to upload archives it does not actually create.

Install only after review. Use it only in a controlled CI/CD environment with trusted packaging plans, path allowlists, secret scanning, and a separate verification step that real archives are created and copied before any deployment treats them as published.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill describes file-reading and file-writing behavior but does not declare permissions, which weakens security review and policy enforcement. In a packaging skill that trusts upstream input and writes artifacts to shared storage, undeclared file capabilities can mask broad filesystem access and make misuse harder to detect.

Tp4

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding
The documented behavior says the skill creates and uploads deployment packages, but the analyzed behavior also computes per-file hashes, generates a detailed manifest, and in live mode may only create the shared directory and emit misleading 'Uploaded' logs without actually uploading artifacts. This mismatch can cause operators or downstream automation to trust nonexistent or incomplete deployments, and the detailed manifest/hash collection may expose sensitive repository structure or file metadata beyond the declared purpose.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill instructs unattended upload of deployment packages to broadly accessible shared storage without any warning, approval gate, or sensitivity checks. Because the skill explicitly trusts upstream packaging plans and skips re-auditing file contents, it can propagate secrets, internal code, or unintended files into a location accessible by all team members and CI/CD systems.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.