The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from user inputs (MOTD, server descriptions, seeds) and external API metadata (Mojang, Fabric, Modrinth). \n- Ingestion points: Asks for user input for server identity fields and queries external APIs via scripts/mcmeta.py. \n- Boundary markers: Absent; user strings like MOTD are written to server.properties or used in commands without explicit delimiters or "ignore" instructions. \n- Capability inventory: Spawns Java processes, manages systemd services, and configures firewall rules via shell commands. \n- Sanitization: The mcmeta.py script uses urllib.parse.quote for API requests; shell command parameters rely on agent-level sanitization. \n- [EXTERNAL_DOWNLOADS]: Fetches Minecraft server jars and mod metadata from well-known official services. \n- Sources: Retrieves assets from Mojang (Microsoft), FabricMC, and Modrinth APIs. \n- Integrity: Implements SHA1 verification for Mojang downloads using the included mcmeta.py script. \n- [COMMAND_EXECUTION]: Executes administrative shell commands for service management and network hardening. \n- Commands: Uses systemctl for service control, ufw for firewall management, and ss for port validation. \n- Safety: Adheres to security principles by instructing the agent not to run as root and to use a dedicated minecraft user. \n- [DYNAMIC_EXECUTION]: Generates systemd service units and command-line arguments for JVM startup at runtime. \n- Pattern: Dynamically constructs ExecStart directives for the minecraft.service based on user-selected RAM allocation and server type.

minecraft-server