text-to-visual

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The picsart install script URL is suspicious because the skill instructs piping an external shell script to bash (curl ... | bash), which can execute arbitrary code; the CDN .webp image is a normal asset and not a download risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The runtime path gen-ai generate / gen-ai batch run in SKILL.md → modes reads the user-supplied “text (paragraph, article, or URL)” and uses it to extract signals and build prompts, so outsider-authored free text is directly ingested.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The SKILL.md prerequisite runs a remote install script at runtime using curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash, which fetches and executes remote code as a required dependency.