product-photo-studio
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill provides a command to download and execute an installation script directly from the vendor's domain:
curl -fsSL https://picsart.com/gen-ai-cli/install.sh | bash. This is a legitimate installer for the skill's required CLI tool, hosted on Picsart's official infrastructure. - [EXTERNAL_DOWNLOADS]: The skill fetches the
gen-aiCLI and associated model metadata from the picsart.com domain. These are official resources provided by the skill author. - [COMMAND_EXECUTION]: The skill relies on executing various shell commands via the
gen-aiCLI, including functions for image generation, batch processing, and pricing estimation. These commands are integral to the skill's intended functionality. - [PROMPT_INJECTION]: The skill ingests user-provided text prompts that are passed to generative models, which constitutes a potential surface for indirect prompt injection.
- Ingestion points: User-supplied strings used with the
-por--promptflags in CLI commands (e.g., inSKILL.mdandreferences/modes/variants.md). - Boundary markers: The instructions do not specify any delimiters or safety markers for separating user-supplied prompts from system instructions.
- Capability inventory: The skill uses the
gen-aiCLI tool, which has the capability to write to the local filesystem and interact with remote Picsart APIs. - Sanitization: There are no documented steps for the agent to sanitize or validate user-provided strings before they are executed in the CLI environment.
Audit Metadata