The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/install_klipper_extras.sh executes shell commands on a remote Klipper printer host using SSH to install various modules and perform system-level updates.
[EXTERNAL_DOWNLOADS]: The scripts/install_klipper_extras.sh script clones third-party GitHub repositories (such as Anonoei/klipper_auto_speed and kyleisah/Klipper-Adaptive-Meshing-Purging) to the remote printer host and executes installation scripts contained within them.
[COMMAND_EXECUTION]: Multiple scripts, including scripts/upload_config.sh and scripts/install_klipper_extras.sh, utilize the Moonraker API to upload and overwrite the printer's primary configuration files (printer.cfg and moonraker.conf).
[REMOTE_CODE_EXECUTION]: The scripts/send_gcode.sh script allows the agent to send arbitrary G-code commands to the printer hardware via the Moonraker API, granting direct control over movements, heaters, and other physical functions.
[COMMAND_EXECUTION]: The scripts/diagnose.sh script performs an analysis by fetching the full printer configuration and executing a Python-based diagnostic pass on the retrieved content.
[PROMPT_INJECTION]: The skill architecture involves fetching and parsing configuration data from a remote printer, which acts as an ingestion point for potentially untrusted data that could influence agent behavior.
Ingestion points: scripts/fetch_klipper_config.sh, scripts/diagnose.sh (reading printer.cfg via the Moonraker API).
Boundary markers: None identified in the parsing logic.
Capability inventory: SSH command execution, G-code execution, and configuration file modification via API uploads.
Sanitization: No specific sanitization or filtering of configuration file content (such as comments) is performed before processing.

3dprint-advisor