hyprland

Fail

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The README.md and install.sh files contain patterns for downloading and executing a remote script directly via shell piping.
  • Evidence: The README.md instructions encourage users to run curl -fsSL https://raw.githubusercontent.com/marceloeatworld/hyprland-ai-skill/main/install.sh | bash for setup across various AI tools.
  • Context: While this is a common developer installation pattern, it facilitates the execution of remote code from the repository at runtime.
  • [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes documentation from an external wiki, creating a surface for indirect instructions to influence agent behavior.
  • Ingestion points: The agent is instructed to read from 40+ markdown files in the references/ directory which are auto-generated from the official Hyprland wiki.
  • Boundary markers: Absent. There are no explicit delimiters or instructions in SKILL.md telling the agent to ignore instructions embedded within the documentation files.
  • Capability inventory: The skill utilizes the hyprctl CLI and hl.exec_cmd function, providing the agent with the ability to execute system commands and modify desktop configurations.
  • Sanitization: Absent. The external wiki content is ingested as raw markdown without filtering.
  • [EXTERNAL_DOWNLOADS]: The skill uses scripts and instructions to fetch content from remote servers.
  • Sources: The scripts/generate-references.sh script clones the hyprwm/hyprland-wiki repository from GitHub. The SKILL.md file also includes links for "Live Fetching" of documentation from raw.githubusercontent.com.
  • Context: These references target official documentation repositories on well-known hosting services.
  • [COMMAND_EXECUTION]: The skill logic and instructions involve executing shell commands on the host system.
  • Evidence: SKILL.md explicitly directs the agent to "Check the user's version with hyprctl version when in doubt."
  • Context: This is intended functionality required to provide version-specific configuration advice for the Hyprland compositor.
Recommendations
  • HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/marceloeatworld/hyprland-ai-skill/main/install.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Nov 12, 2026, 08:32 PM