og-image-design

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the belt CLI tool through Bash commands to perform tasks such as logging in, searching, and running applications.
  • [EXTERNAL_DOWNLOADS]: The skill refers to the installation of the belt CLI via npx and provides links to installation scripts hosted on GitHub. These resources are part of the inference-sh organization, which corresponds to the vendor's infrastructure.
  • [REMOTE_CODE_EXECUTION]: The skill uses the belt app run command to execute remote logic for HTML-to-image rendering and AI image generation. These executions happen within the vendor's inference environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes HTML and CSS strings provided in templates.
  • Ingestion points: HTML and CSS code blocks within SKILL.md templates.
  • Boundary markers: None present to delimit user input from instructions.
  • Capability inventory: The skill is authorized to use Bash(belt *) to interact with remote inference services.
  • Sanitization: No explicit sanitization of user-provided HTML content is performed by the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:25 PM
Security Audit — agent-trust-hub — og-image-design