ascii-art

Fail

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: HIGHPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill's setup instructions direct the agent or user to utilize sudo apt install and sudo snap install for tools like cowsay, boxes, toilet, and ascii-image-converter, granting elevated permissions during installation.
  • [COMMAND_EXECUTION]: The skill relies on shell command execution for its primary functions, including curl for external data retrieval, pip and go for package installation, and direct execution of local binaries.
  • [EXTERNAL_DOWNLOADS]: The skill downloads content and installs tools from several non-trusted external sources, including asciified.thelicato.io, ascii.co.uk, qrenco.de, wttr.in, and the GitHub repository TheZoraiz/ascii-image-converter.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill fetches data from ascii.co.uk, asciified.thelicato.io, and other third-party APIs (SKILL.md).
  • Boundary markers: No delimiters or safety instructions are used when processing the downloaded text before displaying it.
  • Capability inventory: The skill environment supports shell execution (curl, pip, sudo apt), file system writes (/tmp/ascii_art.html), and Python script execution (SKILL.md).
  • Sanitization: Sanitization is limited to basic HTML tag removal using regex, which is insufficient to prevent sophisticated prompt injection or malicious terminal escape sequences from influencing the agent.
Recommendations
  • HIGH: Downloads and executes remote code from: https://ascii.co.uk/art/cat - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 8, 2026, 07:02 PM
Security Audit — agent-trust-hub — ascii-art