The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the terminal tool to install and execute the OpenHands CLI. This is the primary function of the skill, allowing it to delegate complex coding tasks, file modifications, and shell commands within a specified project directory.
[EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to install the openhands package using uv tool install. This fetches the tool from a public registry. The tool is maintained by a well-known open-source project (All-Hands-AI).
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill requires sensitive information such as LLM_API_KEY to function. It correctly instructs the user to provide these via environment variables rather than hardcoding them, which is a standard security best practice for local tools.
[INDIRECT_PROMPT_INJECTION]: As a coding agent that processes project files and user instructions, this skill inherently possesses an attack surface for indirect prompt injection.
Ingestion points: Project files read from the workspace (e.g., source code in src/) and task prompts provided by the user.
Boundary markers: None explicitly implemented in the skill instructions to separate file content from agent instructions.
Capability inventory: Full terminal access and file system modification (via the terminal and file_editor tools within the OpenHands environment).
Sanitization: No specific sanitization of external file content is described in the skill metadata.

openhands