>> is it possibleto filter remote commands, i.e. I want to reject to execute some dangerous commands like "rm"
Seems relatively straightforward - since you are a proxy, then I assume you present your "server" face to the potential clients whose dangerous commands you want to filter. There are actually 2 command "paths" that can pass through your proxy facade: * A single command - as an EXEC channel - easy to handle - you register a CommandFactory and examine the command you are asked to execute. If it is deemed "dangerous" then you can close the channel, or return an error exit code - whatever. Note that this cannot prevent execution of aliases of dangerous commands or scripts that contain dangerous commands in them * As a SHELL channel - this is a bit trickier - what you have to do is to intercept the STDIN and examine the commands before you pass them on to the real shell for execution. The problem here is more complex since the savvy user can ask to execute a script instead of a command, and embed the dangerous commands in the script. Hope this helps.
