The Agent Skills Directory

[PROMPT_INJECTION]: The AGENTS.md file contains explicit behavioral overrides, such as 'Rule 0', which demands absolute obedience to author instructions over following guidelines. A specific note for Codex/GPT-5.2 also instructs the agent to 'fool itself' into thinking it made changes it didn't actually perform, effectively suppressing native agent warnings about external or concurrent file modifications.
[REMOTE_CODE_EXECUTION]: The README.md and SKILL.md files promote an installation method using curl -fsSL https://raw.githubusercontent.com/Dicklesworthstone/slb/main/scripts/install.sh | bash. This pattern involves piping a remote script directly to a shell, which allows for remote code execution.
[COMMAND_EXECUTION]: The core utility of the skill is to authorize and execute shell commands. Files such as internal/core/command.go and internal/core/execute.go utilize os/exec to run arbitrary commands in the user's shell environment, inheriting all current environment variables and credentials.
[COMMAND_EXECUTION]: The installation script and command classification engine include the use of sudo, providing a path for executing commands with elevated privileges during setup or operation.

slb