ClawHub
Back to skill

Security audit

Upstage Document Classification

Security checks across malware telemetry and agentic risk

Overview

This is a coherent document-classification skill that uses Upstage's hosted API, with expected privacy considerations for uploaded documents and local outputs.

Install only if you are comfortable sending the documents or document URLs you classify to Upstage using your Upstage API key. Avoid regulated, confidential, or highly sensitive documents unless your organization has approved the service's privacy, retention, billing, and compliance terms, and remove local classified JSON or split-PDF outputs when they are no longer needed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (6)

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill explicitly instructs writing classification results and split document outputs to local temp paths, but provides no guidance on handling potentially sensitive document contents, retention, access controls, or cleanup. Because this skill processes user documents that may contain PII, contracts, invoices, or resumes, persisting outputs to disk can expose sensitive data to other local users, logs, backups, or later processes.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The examples instruct users to send PDF URLs and potentially sensitive document identifiers to a third-party API without any privacy, consent, retention, or data-classification warning. In a document-processing skill, this increases the risk of inadvertent disclosure of invoices, contracts, receipts, and other regulated or confidential business data.

Missing User Warnings

Low
Confidence
74% confidence
Finding
Referencing an API key from an environment variable is standard practice, but the documentation omits basic credential-handling guidance. In example code intended for broad reuse, failing to warn against logging, hardcoding, or exposing the key can contribute to accidental credential leakage.

External Transmission

Medium
Category
Data Exfiltration
Content
---

## curl Example (Single-Document Classification)

```bash
curl -X POST "https://api.upstage.ai/v1/document-classification" \
Confidence
89% confidence
Finding
This example explicitly transmits user-supplied document data to an external API endpoint. In the context of document classification, that behavior is expected, but it is still security-relevant because it can move sensitive business or personal documents outside the local trust boundary.

External Transmission

Medium
Category
Data Exfiltration
Content
client = OpenAI(
    api_key=os.environ["UPSTAGE_API_KEY"],
    base_url="https://api.upstage.ai/v1/document-classification"
)

response = client.chat.completions.create(
Confidence
90% confidence
Finding
The configured base URL points to a remote third-party classification service, so all submitted document content is sent off-host for processing. That is inherent to the skill's function, but without prominent disclosure it can surprise users and create confidentiality or compliance issues.

External Transmission

Medium
Category
Data Exfiltration
Content
## curl Example (Single-Document Classification)

```bash
curl -X POST "https://api.upstage.ai/v1/document-classification" \
  -H "Authorization: Bearer $UPSTAGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
Confidence
89% confidence
Finding
The curl example demonstrates direct external transmission of potentially sensitive document references and authorization credentials to a remote API. While this is intended functionality, it still warrants treatment as a security concern because misuse can leak confidential document access patterns or contents.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.