ClawHub
Back to skill

Security audit

Workflow Orchestrator Free

Security checks across malware telemetry and agentic risk

Overview

This skill is a local workflow-building guide whose file writes, shell execution, and API examples fit its stated automation purpose, though users should review generated scripts and stored data carefully.

Install only if you are comfortable with an agent creating and running Bash workflow scripts. Review any generated run.sh and any sourced connection files before execution, keep API credentials in environment variables or a keychain, and avoid storing sensitive API responses in workflow data or logs unless that is intended.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
The skill explicitly encourages `source workflows/components/connections/api-auth.sh`, which executes arbitrary shell code from a reusable component path rather than treating connection data as inert configuration. In an agent skill with `exec` capability, this broadens the trust boundary and can enable hidden command execution, credential exfiltration, or environment tampering if a connection script is modified or replaced.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill persistently writes workflow state, data, logs, scripts, and outputs to disk but does not prominently warn users that files will be created or overwritten. In an agent context, silent persistence can expose sensitive data locally, overwrite existing artifacts, and make later unintended reuse or leakage more likely.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The examples perform outbound `curl` requests to external APIs and save returned data locally without any privacy, egress, or trust warning. In an agent environment, users may adapt these examples to real endpoints containing credentials or sensitive business data, creating risk of unintended transmission and local persistence.

External Transmission

Medium
Category
Data Exfiltration
Content
flock -n 200 || { echo "另一个实例正在运行"; exit 0; }

# 节点1：获取数据
curl -s "https://api.example.com/data" > data/01-fetch.json

# 节点2：过滤数据
jq '[.[] | select(.status=="active")]' data/01-fetch.json > data/02-filter.json
Confidence
88% confidence
Finding
The skill includes explicit external transmission via `curl` to an HTTPS endpoint and persists the response. External network egress is contextually aligned with ETL/reporting workflows, but it still becomes a real security concern when used with live endpoints because it can transmit sensitive identifiers, metadata, or tokens and normalize network access without user review.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.