emacs-agent-assisted-second-brain
Table of Contents
lsp contributions to emacs notes
[ ]when I am editing a prompt, I should be able to get inline suggestions and additional resources from an agent via lsp[ ]when I am editing my notes, I should be able to get inline resource provided by an agent via lsp[ ]when I am editing a prompt it should create a buffer in an emacs server that is accessible from other emacs clients so that I can easily transfer content from my other emacs notes
agentic actions based on notes interactive engagement with second brain
[ ]when I am engaging with my second brain (emacs org-mode w/ org-roam), I should be able to get dynamic summaries from agenda view and from an agent leveraging the roam-db on action plans based on recency, relevancy to my recent thought additions, or relevancy to upcoming scheduled items[ ]when I update my notes, backend async long-running agents should create actions plans based on my notes. These action plans should be embedded in the same org-roam text files and db file and should use automerge to avoid conflicting writes[ ]when an agent creates an action plan, it should be able to notify me and solicit approval via various channels (SMS, webapp push notification, passive tagging of notes with referenced generated action plan)[ ]given an action plan is approved for overall execution, when the agent requires subsequent tool approval, it should be able to use similar channels to get approval[ ]given an action plan is approved for overall execution, when the agent requires subsequent tool approval, if I do not approve within a period of time, it should persist the outstanding action plans in a queue for later review
diverse surfaces for interaction
[ ]I should be able to update my notes from various clients to be synced to a centralized backend (webdav) that is visible to my agents[ ]I should be able to pull summarize based on my second brain backend (open claw?), including history of changes (eg: git-managed code changes)
agentic usage for task execution
[ ]when i am iterating with an agent to complete a task on remote server, I should be able to also access the agent conversation via emacs tramp buffer on a local laptop machine