The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides tools (pty_spawn, pty_write) for executing and interacting with arbitrary shell commands. This functionality is the primary intended purpose of the skill, enabling development workflows that require background processes and interactive terminal control.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by design.
Ingestion points: The pty_read tool in SKILL.md retrieves output from external processes and provides it to the agent.
Boundary markers: None are specified in the tool documentation to distinguish process output from system instructions.
Capability inventory: The agent possesses command execution and interaction capabilities via pty_spawn, pty_write, and pty_kill.
Sanitization: No sanitization of process output is described; the agent receives the raw (or optionally regex-filtered) output of the spawned process.

opencode-pty