ClawHub
Back to skill

Security audit

orshot

Security checks across malware telemetry and agentic risk

Overview

The skill is legitimate Orshot API documentation, but it includes live social publishing and destructive remote actions without strong user-confirmation guidance.

Install only if you intend to let the agent work with Orshot workspace APIs. Use least-privilege API keys or OAuth scopes, prefer drafts for social posts, and require explicit confirmation before publishing, deleting templates, or deleting brand assets.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Description-Behavior Mismatch

Medium
Confidence
96% confidence
Finding
The manifest presents the skill as visual-content generation, but this section also enables direct publication to third-party social accounts. That expands the skill from content generation into external action execution, creating a capability mismatch that could cause an agent or user to trigger real-world posts without expecting that side effect.

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The skill metadata frames Orshot as a rendering API, but this section additionally describes OAuth developer app creation and third-party workspace access. That broadens scope into delegated access and credentialed cross-workspace operations, which is more sensitive than simple media generation and should be declared up front.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The social publishing instructions describe posting to external platforms, including immediate publication when isDraft is false, but they do not prominently warn about irreversible external posting consequences. In an agent setting, that omission increases the risk of accidental posting, brand damage, or leakage of sensitive/generated content to public accounts.

Missing User Warnings

Low
Confidence
86% confidence
Finding
The skill documents DELETE endpoints for templates and brand assets without emphasizing that these actions remove resources and may be irreversible. While common in API docs, the lack of caution is riskier in an agent-executed context where destructive actions may be triggered from ambiguous prompts.

External Transmission

Medium
Category
Data Exfiltration
Content
### 6. Delete Studio Template

**DELETE** `https://api.orshot.com/v1/studio/templates/:templateId`

### 7. Duplicate Studio Template
Confidence
83% confidence
Finding
https://api.orshot.com/

External Transmission

Medium
Category
Data Exfiltration
Content
**GET** `https://api.orshot.com/v1/brand-assets`

#### Delete Brand Asset
**DELETE** `https://api.orshot.com/v1/brand-assets/:assetId`

### 10. Enterprise API Endpoints
Confidence
82% confidence
Finding
https://api.orshot.com/

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
SKILL.md:570