Automation

Ships a commit skill and holds every Bash git commit to it: a commit the skill did not open, a blanket git add, a signature trailer, a secret, an ignored path, or a push or branch change nobody asked for is stopped before git runs.

Measures how much context each Bash command spent and tells the model, past a size limit, which narrower command would have answered the same question.

For mod development: runs a live check of plugins in a fresh Claude Code session in tmux, types the steps one by one, collects the pane and the transcript, then deletes the temp directory, its transcripts and the inline store files.

Writes a prompt longer than the limit (2000 characters by default) to a temp file and sends the model its first 200 characters with the path, so one paste does not fill the context.

Runs slash commands joined with && one after another, waits for each command's turn, pane or dialog to end, and stops the chain at a command that failed or a turn that ended without an answer.

After an edit that builds SQL by joining or interpolating strings, names each line, so the model passes the values as query parameters.

Paints Markdown onto Claude Code's prompt box as you type: fenced code becomes a syntax-highlighted card (even before you close the fence), and inline code, bold, italic, lists, task lists, quotes, tables, links and headings are styled. Paint only, your text is never changed.

State-aware Neovim diff panes

One-pass local response cleanup without agent retries

Compacts when the work moves on from one sub-task to the next instead of when the context window fills, including in the middle of a long turn. Once the context is past a floor, a model judges each step, reading what you asked for and where Claude has got to but never any tool output, for whether the work moves on from a finished piece to another; if it does, taskcut runs Claude Code's own...

Push guard: before a git push to a public GitHub repo (or making one public) it scans what would go out for Supabase project IDs, keys, tokens, emails, local paths, private repo names and .env files, and asks in a one-key popup. /pushor toggles. Hungarian or English.

A pixel-art office in a local web page: every Claude session gets a desk, walks to the filing cabinet when it reads, types when it edits, flies paper planes when it messages, and shows a live Done-by clock

An opinionated engineering workflow for Claude Code: brainstorm -> plan -> TDD execution with fresh-context subagents -> verification -> captured learnings. Lean reviewers, stack-specific patterns for TypeScript/React, Python and Flutter, and six light safety hooks.

Compacts the main conversation once it has sat idle for idleMinutes, while the prompt cache is still warm, and hands every compaction of the main conversation (manual, auto, plugin, precompute) the Compact Instructions section of the user's AGENTS.md. Skips while a turn runs, below minMessageTokens of conversation, and past cutoffMinutes (a wake from sleep, when the cache has likely expired)...

Answers trivial questions (what branch, what changed, did the last test pass) from git and the last test run on your machine, with no model tokens.

Runs long Claude Code jobs overnight within your usage limits: pauses just before the 5-hour or weekly limit, resumes the same session after the reset (or relaunches it days later) and works through a list of jobs you hand it with /noctis:start . Also runs in Codex CLI, Antigravity CLI, Droid and Copilot CLI.

Jev Gate in one plugin: the orchestration gate, the extractive compactor, the Jev effort and model router and the jev_evidence tool. Every part is on by default; each has its own on/off option.

/mejora y una frase: Claude te hace 3-5 preguntas, escribe un prompt con contexto (sin rol), lo puntúa y lo corrige hasta el 9/10, con bucle o sin bucle. Método de diariovida.com.

Experimental Claude Code mods for GraphOS Agent Services. GraphOS Inspector explains each execute call in a docked pane before you approve it, shows what came back with clickable records, and can auto-approve the reads you write trust rules for.

Before a shell command runs, parses it with the shell's own parser and refuses one that does not parse, such as an unclosed quote or heredoc, so none of it runs.

Adds Exa's results to WebSearch, and saves each WebFetch page to disk with the lines that answer the prompt

Prevents hand-editing lockfiles; they must change through the package manager.

Blocks commands that upload files to external services like pastebins and file-sharing sites.

Allows writes only inside the project root, resolving symlinks and .. tricks.