Docs: Clarify run_terminal_command and add execute_internal_command

Updated the description for `run_terminal_command` to specify it runs in an "empty" Docker sandbox.
Added documentation for the new `execute_internal_command` tool, which executes commands on the host machine.
This commit is contained in:
Slipstream 2025-05-27 22:35:27 -06:00
parent 64857c683f
commit bb8cdadf47
Signed by: slipstream
GPG Key ID: 13E498CE010AC6FD

View File

@ -151,7 +151,8 @@ OS: Arch Linux x86_64; Host: 1.0; Kernel: 6.14.5-arch1-1; Uptime: 11 days, 21 ho
- `calculate`: Evaluate math expression. Ex: `calculate(expression="...")`.
- `run_python_code`: Execute simple, safe Python code sandbox. USE CAUTIOUSLY. Ex: `run_python_code(code="...")`.
- `create_poll`: Make a poll message. Ex: `create_poll(question="...", options=["...", "..."])`.
- `run_terminal_command`: Execute shell command in Docker sandbox. EXTREME CAUTION. Avoid if unsure. Ex: `run_terminal_command(command="...")`.
- `run_terminal_command`: Execute shell command in a (empty) Docker sandbox. EXTREME CAUTION. Avoid if unsure. Ex: `run_terminal_command(command="...")`.
- `execute_internal_command`: Execute shell command on host machine. EXTREME CAUTION. Avoid if unsure.
- `get_user_id`: Find user ID from username/display name. Ex: `get_user_id(user_name="...")`.
- `no_operation`: **MUST call this after all other necessary tool calls are done.** Use immediately if no tools needed. Does nothing itself.