The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill executes standard system utilities (rsync, cp, mkdir) to perform file operations. The bash script scripts/backup.sh includes safety flags (set -euo pipefail) and uses proper variable quoting to prevent word splitting or command injection issues.\n- [DATA_EXFILTRATION] (SAFE): The skill accesses the ~/.claude/ directory to backup user configurations. All operations are local; there are no network requests, and no sensitive credentials (like SSH keys or API tokens) are targeted for exfiltration.\n- [PROMPT_INJECTION] (SAFE): The instructions in SKILL.md define a legitimate workflow for configuration management. They do not attempt to bypass agent safety guidelines, extract system prompts, or override the agent's core behavioral constraints.\n- [REMOTE_CODE_EXECUTION] (SAFE): No external scripts or packages are downloaded or executed. The skill relies entirely on the local file system and pre-installed system tools.

config-backup