The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to install the Coder CLI using a shell pipe (curl -fsSL https://coder.com/install.sh | sh). This pattern executes a remote script directly with shell privileges. While the source is the official vendor domain, this method bypasses standard package verification.
[EXTERNAL_DOWNLOADS]: The skill fetches an installation script and documentation from the author's official website (coder.com) to set up and guide CLI operations.
[COMMAND_EXECUTION]: The skill relies on the coder CLI to perform all workspace operations. This includes running arbitrary shell commands inside remote development environments via coder ssh <workspace> -- <command>.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data retrieved from the Coder deployment, creating a potential vector for malicious content to influence the agent.
Ingestion points: The agent reads build logs (coder logs) and file contents (coder ssh cat) from remote workspaces.
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands when processing this external data.
Capability inventory: The skill possesses significant capabilities, including workspace lifecycle management (creation and deletion), file writing via tee, and network port forwarding.
Sanitization: No sanitization or validation steps are defined for the logs or file contents retrieved from the workspaces before the agent acts on them.
[PERSISTENCE]: The skill uses coder config-ssh -y to modify the local ~/.ssh/config file. While this is the intended functionality for integrating SSH-aware tools with the workspaces, it involves automated modification of sensitive system configuration files.

coder-workspaces