/less-permission-prompts <-- this skill scans your history for things that are well-known/safe commands that previously called for you to act y/n on. Big time saver, and a good bridge between --dangerously-skip-permissions and "OMG YES hwo many times do i have to approve this"
/recap <-- The anthropic docs say this is to invoke a session recap, without any context as to why you would do so. I can see this as a good tool for context management outside of Claude Code. Write this out to MD for your next agent, or pass it as a stop hook to a in project memory file. This gives you a brief of what we did/what's next in a few sentences.
/Advisor <-- allows you to run Sonnet, then invoke your "advisor" agent when Sonnet gets off track. Interesting play if you primarily drive Sonnet and then want to appropriate some of your tokens to a more powerful model.
/Dashboard <-- Spawns a remote session that designs a dashboard for your data sources. Wild - I haven't tried this yet - has anyone used this one yet?
[link] [comments]
