Tools & agents

Keeps a glossary of the conversation's terms and a map of where the work stands, in panes opened from a band above the prompt, with a btw input for side questions.

PIANO-esque cognitive layer: persona-keyed state, heartbeat liveness, controller with classify/actuate

Every look you reject becomes a rule for that project, with your screenshot, so Claude never brings it back; it also says when Claude changed UI without a preview you approved

Codex inside Claude Code: it picks up the chat when your Claude limit runs out, talks with you through Codex's live voice, makes images Claude can ask for, and browses in Claude's browser pane

Ather Automata: what needs you and what is next in every S2 session (Plan, Build, Prove, Ship), a newcomer tour, autonomy windows, and guards that catch false build passes and lost merge edits.

Hands a long session off cleanly: at a context threshold Claude updates the docs and writes a handoff, the session compacts with its instructions, then a kickoff prompt carries on. Gives Claude a live context reading and a /wrap-up skill

Live plan progress bars above the Claude Code prompt: stages, steps, a pixel fill and soft sounds for decision, error and done

ND-friendly mods for Claude Code: patch status line and friends.

Live plan, % done, time left and a one-line 'now' for long tasks, in a pane and on the working row.

An inbox for review requests (oldest first, with AI summary, risk and release impact) and your own PRs (what needs action first), with explain and approve from the pane

Keeps Claude's todo list in a side pane beside the transcript, open only while the list has items.

Automated multi-agent code review cycle. Spawns reviewers in parallel (plus Codex when available), applies fixes per CLAUDE.md policy, admits a commit only when a reviewer saw exactly what it records, and asks before a push, pull request, merge or release you did not request, unless you let it go further.

A live side pane that maps a long Claude Code session: goal, current path, possible detours, decisions, open questions, checkpoints, working files and what to resume next. It observes on its own and asks before changing your intent.

doc-desk Mod: Claude が質問票 JSON を open_form ツールに渡すと、ブラウザに一枚のフォームを出して人に答えてもらい、回答を【doc-desk 回答】の固定形にして Claude に届ける (waitSeconds 以内ならツールの結果として、過ぎれば user turn として)。書き上げた文書を open_review に渡すと、人が段落や文字列に付けた指摘を同じ固定形で届ける (指摘モード)。

Claude orchestrates, Command Code (cmdc) implements: a tool that hands a task to cmdc and returns its diff for review, with a live pane of the run.

Keeps raw tool output out of the context: sandboxed execution, a local FTS5 index, routing hooks and session resume, with strands-decider making the close calls.

A podcast player inside Claude Code: subscribe to feeds, browse episodes in a pane, and play them with a now-playing bar above the prompt

A per-project progress board: the main model records each conversation's work as nodes through the progress tool, kept under ~/.claude/progress/ or committed with the project; an optional Artifact canvas mirrors it

A per-project backlog of items needing your attention, parked by you or the orchestrator

SDD estilo zero-pi dentro de Claude Code, con un modelo del CPAM por fase

Opt-in persistent task checklist: the agent asks to use it, the user approves, checks items off and alone can clear or disable it.

A pomodoro timer on the prompt's hint line whose break lands while Claude works, with buttons beside it, a history and a report: /pomodoro start

Runs 2-4 attempts at a task in parallel git worktrees, each with a different approach; compare their diffs and adopt the best.

Finds the commit that broke something: git bisect with your test command in a temporary worktree (your working tree is never touched), live progress in a pane, the culprit with its diff stat handed to Claude.