The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the creation and execution of shell commands and scripts. It implements safety-first rules, such as mandatory double-quoting of variables to prevent word splitting, explicit error handling using 'set -euo pipefail', and the use of 'printf' for predictable output. These measures are designed to ensure robustness and prevent common shell-based vulnerabilities.\n- [PERSISTENCE]: The skill provides specific templates and guidance for modifying shell profile files (e.g., '.bashrc', '.zshrc') to configure interactive features like tmux autostart for SSH sessions. These persistence mechanisms are within the intended scope of dotfile management and include environment guards (e.g., TTY checks and SSH_CONNECTION detection) to prevent hijacking non-interactive automation or file transfers.\n- [INDIRECT_PROMPT_INJECTION]: As an auditing tool, the skill is designed to process existing shell scripts provided by users. This creates a surface for indirect prompt injection. However, the skill mitigates this risk through a rigorous verification checklist that requires the agent to use static analysis tools like ShellCheck and to validate shell features against specific versions before execution.

shell-scripting