Back to skill

Security audit

Orshot Design: on-brand template design rules for images, ads, thumbnails and video

Security checks for vulnerabilities and agentic risk

Overview

The skill’s hosted design workflow fits its purpose, with ordinary considerations around sharing content, editing templates, and rendering costs.

Use the intended Orshot workspace and share only content you are permitted to upload. Preserve a copy before refining an existing template. Validation renders may consume credits.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

The workflow sends user briefs, copy, and assets to hosted Orshot tools without a privacy warning.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The workflow instructs the agent to submit the user's brief and job description to orshot_plan_design, then finished copy and real assets to orshot_create_template_design. Although the file identifies the hosted Orshot MCP server, it does not warn that these steps transmit user content to a third-party service, potentially including confidential or unpublished material.

Content

No source excerpt is available for this finding.

The refinement workflow directly invokes rendering despite the manifest explicitly excluding it.

Low
Category
Excessive Agency
Confidence
98% confidence
Finding

The manifest at L03 states that rendering belongs to the separate orshot skill, and L25 repeats that boundary. However, L85 explicitly directs this skill to call orshot_render_template; L53 and L69 also require render checks. This goes beyond the thumbnail-based create-and-critique workflow declared for this skill.

Content

No source excerpt is available for this finding.