Skills
skills/picsart/gen-ai-skills/text-to-visual/Gen Agent Trust Hub

text-to-visual

Fail

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to download an installation script from https://picsart.com/gen-ai-cli/install.sh, which is an official resource belonging to the skill's author.
  • [REMOTE_CODE_EXECUTION]: The installation process involves piping a remote shell script directly into bash (curl -fsSL ... | bash). This is a standard installation pattern for the vendor's command-line tool.
  • [COMMAND_EXECUTION]: The skill utilizes several local shell utilities for workflow automation, including gen-ai, jq, pbpaste, xclip, xargs, curl, and open.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it ingests untrusted data from URLs and user text for prompt construction.
  • Ingestion points: Content is ingested from user-provided paragraphs, articles, and URLs in SKILL.md and article-set.md.
  • Boundary markers: There are no explicit delimiters or system instructions to prevent the model from obeying instructions embedded within the source text.
  • Capability inventory: The gen-ai CLI tool provides network access (to Picsart Drive/CDN) and file system write capabilities.
  • Sanitization: The instructions do not describe any sanitization or filtering of the input text before it is used to generate visual prompts.
Recommendations
  • HIGH: Downloads and executes remote code from: https://picsart.com/gen-ai-cli/install.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 25, 2026, 12:52 AM
Security Audit — agent-trust-hub — text-to-visual