The Agent Skills Directory

COMMAND_EXECUTION (SAFE): The skill includes an installation script (scripts/install_statusline.sh) and a generation script (scripts/generate_statusline.sh). These are used to configure the agent's terminal UI and run periodically to update information. Both scripts use standard shell commands and do not execute untrusted remote code.\n- EXTERNAL_DOWNLOADS (SAFE): The documentation references the @anthropic-ai/ccusage package for cost tracking. This utility is produced by a trusted organization (Anthropic) and is not automatically installed or executed by the skill's scripts without user intervention.\n- DATA_EXPOSURE (SAFE): The scripts access the current working directory, username, and local configuration files (~/.claude/settings.json) to populate the statusline. This data is used only for local display and is not exfiltrated via the network.\n- PROMPT_INJECTION (SAFE): The skill handles environment data such as model names and file paths. These inputs are used for terminal rendering and do not influence the agent's core logic or bypass safety filters.

statusline-generator