Just saw this - Obsidian is now doing automated security scanning for every community plugin, and publishing scan results: The future of Obsidian plugins - Obsidian
Importantly, they’re now (automatically) reviewing every released version; before, they would only do an initial manual review upon submission to the community plugin repo.
This is super good news! But would the automated scanning have prevented this attack? The Shell Commands plugin is legit; it wasn’t backdoored or compromised. So unless they pair this with a policy change about code execution, I think the attack pattern remains intact.
Ah yep, I think that’s the case! It is extremely unfortunate that the auto-trigger and the “turn confirmation off” are part of the intended functionality of this plugin. The Obsidian post did mention that they are including capability disclosures as part of the review, but it seems like they’re not scanning those, only relying on the plugin authors to declare.