The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads existing application configuration and uses that data to generate restoration commands, creating a surface where malicious data in the configuration file could influence agent actions.
Ingestion points: Existing configuration values are read from ~/.config/cmux/cmux.json using the cmux-settings get command as specified in Step 3 of the Workflow in SKILL.md.
Boundary markers: The instructions do not define delimiters or provide warnings to the agent to ignore instructions embedded within the retrieved configuration data.
Capability inventory: The skill has the capability to write to the filesystem and execute shell commands via the cmux-settings helper script across all defined workflows.
Sanitization: No sanitization, escaping, or validation logic is applied to the data retrieved from the configuration file before it is interpolated into shell commands for the reporting and restoration steps.
[COMMAND_EXECUTION]: The skill executes shell commands to interact with the local environment and application settings.
Evidence: The SKILL.md file contains logic to locate and execute a local helper script cmux-settings to perform get, set, unset, and validate operations on the application's JSON configuration.

cmux-keyboard-shortcuts