The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches the latest Cursor AppImage from https://download.cursor.com/linux/appImage/x64. This is the official distribution domain for the Cursor IDE.
[COMMAND_EXECUTION]: Utilizes shell commands for system-level tasks including package management via brew and winget, file operations for backups, and extension management using the cursor CLI.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes local configuration files which could serve as a vector for indirect injection if those files contain malicious content.
Ingestion points: Reads extensions-backup.txt, settings.json, and keybindings.json as seen in SKILL.md and references/migration-between-machines.md.
Boundary markers: Absent. The skill does not implement delimiters or warnings to ignore instructions embedded within the processed configuration data.
Capability inventory: The skill is granted Bash(cmd:*), Read, Write, and Edit permissions, allowing it to execute commands based on file content.
Sanitization: Absent. Data from local files is passed directly to shell commands (e.g., xargs or while read loops) without validation or escaping.

cursor-upgrade-migration