list-directory

The list-directory skill provides a fundamental mechanism for agents to interface with the local or containerized filesystem. It is designed for tasks requiring visibility into the workspace, such as identifying project dependencies, locating source files, or confirming the successful execution of preceding operations like file creation or download. By providing structured metadata including file names, types, sizes, and last modification timestamps, this skill empowers agents to make informed decisions about subsequent actions.

• File and directory discovery: recursively or directly view contents within any accessible path in the agent's sandbox environment.

• Metadata retrieval: accurately capture file attributes such as size and last modification time to help the agent track changes or identify the most recent versions of documents.

• Contextual workspace navigation: essential for multi-agent workflows where agents need to audit the filesystem state before initiating complex tasks like build automation, batch processing, or data parsing.

• Security-first integration: operates within OpenAkita's multi-layer sandbox framework, ensuring directory listing is confined to authorized paths and prevents unauthorized system-wide traversal.

• Usage: Pass a valid string path as the primary argument. Use '.' for the current directory or provide an absolute/relative path to explore other areas.

• Best practices: Always verify if a file exists using this skill before attempting read-file or write-file operations to prevent runtime errors. When working with large directories, note that returning high volumes of metadata may impact agent memory usage. This skill is a prerequisite for file-intensive tasks and integrates seamlessly with other tools like search, read-file, and write-file to form a complete file management pipeline.