The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill explicitly directs the agent to retrieve plaintext passwords from ~/.config/remote-shell/hosts.json and use them in shell commands via the sshpass utility. Providing credentials as command-line arguments is a high-risk practice as they become visible to other users and monitoring tools in the system's process list.
[COMMAND_EXECUTION]: The skill's primary function is the execution of arbitrary commands on remote systems. The fallback mechanism utilizes ssh with the -o StrictHostKeyChecking=no flag, which disables host key verification and exposes the connection to potential Man-in-the-Middle (MITM) attacks.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and displays output from commands executed on remote servers, which could be exploited by an attacker to inject malicious instructions into the agent's context.
Ingestion points: Output from remote commands executed via remote-shell or ssh (described in SKILL.md).
Boundary markers: None; the skill does not use delimiters or provide instructions to treat remote output as untrusted content.
Capability inventory: Full shell command execution on remote hosts, local command execution via sshpass, and file transfer capabilities (SFTP).
Sanitization: There is no filtering or validation performed on the data returned from remote hosts before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill suggests the installation of the remote-shell-cli package from the npm registry.

remote-shell