write-upl-prompts

Pass

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references several download links for pre-compiled binaries of the 'upl' CLI tool hosted on the author's GitHub repository (github.com/DavidValin/universal-prompt-language). These downloads are consistent with the skill's primary purpose of providing a development environment for UPL prompts.
  • [COMMAND_EXECUTION]: The documentation instructs the user and agent on how to install the downloaded binaries using system commands such as chmod +x, sudo install, and uname -m. These are typical operations for software installation on Linux, macOS, and Windows systems.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-defined prompt requirements and existing .upl files, which are untrusted data sources.
  • Ingestion points: User requests to create, validate, or fix .upl files (SKILL.md).
  • Boundary markers: The skill enforces strict structural delimiters including the metadata delimiter (--), heredoc tags (>>>, <<<), and specific expansion tags ([[[, {{{).
  • Capability inventory: The skill uses commands to install and run the 'upl' binary, as well as file writing operations to save prompts.
  • Sanitization: The skill includes a detailed validation checklist to ensure output conforms to a strict grammar, reducing the risk of accidental instruction following by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 11, 2026, 05:34 PM
Security Audit — agent-trust-hub — write-upl-prompts