vault-jot
claudeCapture ideas, takes, and reading items into an Obsidian vault inbox, and incubate ideas from a pane
Claude Code and Codex plugins their authors filed under "productivity". All categories.
Capture ideas, takes, and reading items into an Obsidian vault inbox, and incubate ideas from a pane
Checks the agent's final reply in each turn against the project's writing rules with TypeSafe Jev.
Worktrees, services and Claude's own processes in one pane
Renders the Mermaid diagrams Claude Code writes or reads as elegant cards in the transcript: colored Unicode art in any terminal, real pictures in kitty-graphics terminals such as Ghostty and herdr, native SVG on desktop.
A usage board for Claude Code: a weekly pace gauge and a sliding ticker of your spend per model, in the terminal and on the desktop.
A pane with a live SVG map of what Claude is doing: the turn's phase, the last tool, the streamed thought, and every file it touched
Spend Claude Code tokens where they count: moves a new task into a fresh session (a new terminal window, or a git worktree of its own for unrelated work) instead of dragging a long conversation along, and runs small, clear tasks on Sonnet. Works in tmux, Linux, macOS and Windows terminals or your own terminal command; Superset via a setting. -41% per session in its benchmark; your prompts go out as typed.
Install all 11 Claude Code mods. Includes optional self-drive, MCP controls, side chat and session tools.
Read-only side chat: /aside opens a pane beside the transcript where you ask about the session so far; answers come from a tool-less fork of the session's own transcript and nothing is written back into the main thread.
/changes: the session's edits grouped by concept, then files, then the exact code; point at lines and ask Claude about them or have it change them
No description.
No description.
No description.
Opt-in mod for /review-pr: live board (seats, build gate, cost), per-seat tool-call budget, no general-purpose seats, and a dry-run posting guard.
A heads-up display for the session: the questions Claude asked that you haven't answered, one-line answers to the questions you asked, every link and file it handed you, what is still running, heads-ups and the next step. A small model reads each finished turn. /hud opens the pane.
Collection of skills I use in my daily agentic workflows
Know when Claude needs you or finished something long: a desktop notification, a sound, a toast, a band and an /lk-attention inbox. Linux with GNOME.
Don't lose a session to the context limit: watches context fill, writes a handoff into .claude/brain/ before the limit, then clears and resumes from it.
One line above the prompt with the repo, branch and worktree, how dirty it is, and the account's usage windows and prompt cache, plus a /lk-where pane.
Select text in the conversation, then quote it into your prompt (it stays outlined), ask about it or turn it into a task.
See your subagents: an expandable live tree of every subagent, a timeline, Agent-row cards and a band above the prompt
A band above the prompt for a beads task queue: the ready count when idle, the task this session claimed while working
Panes the model can draw (markdown, stats, tables, SVG, buttons) and media curation for audio, images, video and documents
Records what Claude Code spends, message by message, and shows where it went: on which feature, and how much of it is on features at all.
Ops board mod for skillator: a pane of the TICKETS.md board, a band above the prompt with machine and usage facts, and /list-tickets, /pending, /ops commands.
One table for the session: name, context, 5-hour and weekly quota with resets, subagents, tokens and cost; feeds claude-sabha's status files.
Switches a running session between api.anthropic.com and the local Model Gateway: onto the gateway once Remote Control has registered, on /gateway on, or as each session starts after /gateway setup, and back to api.anthropic.com whenever the gateway stops answering.
Collection of skills I use in my daily agentic workflows
Type >> on and messages sent while Claude works wait until the current task is finished, in an editable queue above the prompt; >> off is back to normal, and >> lines never reach Claude's history. /now <message> interrupts and sends at once.
Claude's plan in a local browser board you can edit while it works (micro-steps, skip, reorder, stop; finished sections locked), with its progress above the prompt, its subagents and background tasks, and session statistics.