ClaudeCodeMod

All shelves / Mods / Workflow

agent-status

By dodobrands

agent-status is a Claude Code workflow mod by dodobrands.

In the desktop app, puts an icon before each session's name for working, waiting on you, stuck, or done, and lists its background tasks above the prompt.

Draws

Install

The marketplace name is ai-hub.

In your shell
claude plugin install agent-status@ai-hub --scope user
Open the repo

It runs inside the Claude Code process. The side panel is only what the author's README names. Read the repo before you install, or run claude plugin validate on the plugin folder.

Version
0.10.0
License
MIT
Repo stars
8
Last change
2026-10-20

What it can reach

  • Draws in the session
  • Sees tool calls

The README is in Russian. Add the ai-hub marketplace first; the repository README gives claude /plugin marketplace add dodobrands/ai-hub. In the app: + by the prompt box, then Plugins, Add plugin, and agent-status from ai-hub. Then allow four tools in ~/.claude/settings.json, mcp__ccd_session_mgmt__get_session, mcp__ccd_session_mgmt__set_session_title, mcp__ccd_sidebar__list_sections, and mcp__ccd_sidebar__move_sessions, so the mod's calls skip auto mode's classifier and ask nothing. It loads in new sessions, and in open ones after /reload-plugins or a restart of the app. The icons and moves need the desktop app; in the terminal only the task list and the button remain. The openSection option moves a session you write in to that sidebar section, off by default, and sections whose name starts with keepPrefix, 📌 by default, are left alone. A title you set by hand changes only after you confirm. /agent-status-usage shows how many Haiku tokens it used today.

The marketplace, ai-hub, holds Dodo Brands' 14 AI Hub plugins, spike, kaiten, time, and discovery among them. This one installs on its own. Seeing which of many parallel sessions needs you is why this page files it under workflow. peek-vedantandhale opens a pane of background shells and subagents, with elapsed time and their last output lines. whats-agent-doing draws a box above the prompt that names the step Claude is on and each background agent. This one marks each session in the desktop app's sidebar, so a dozen agents can be read at a glance. Haiku decides waiting, stuck, or done from the reply's text, through Claude Code's own model client, at about 500 to 700 tokens a turn; the other icons use no model. A background wait that names a Databricks run or gh run watch gets a link to that run, and a done session shows a Task done row with a button that clears the check.

Demo

The opening of the agent-status README, written in Russian: a table of the icon put at the start of each session's name in the desktop sidebar, for working, a background task running, waiting for your answer, stuck, and done; a note that Haiku through $.model.complete picks the last three; and the list of background tasks above the prompt.
Screenshot of the author's GitHub page.

What it hooks into

  • session.startWhen a session starts
  • command.runWhen you run its slash command/agent-status-usage
  • tool.callAround a tool call, where it can hold one
  • prompt.submitWhen you send a prompt
  • turn.completeWhen a turn ends
  • ui.renderDraws into the sessionAbovePrompt

Read from the on() calls in its hook modules. Open its source under Files to check what each one does.

Slash commands 1

  • /agent-status-usageСколько токенов Haiku мод agent-status потратил сегодня

Files

The author's README and source, 4 files. They load when you open them.

Facts

Author
dodobrands
Category
Workflow
Marketplace
ai-hub
Repo
dodobrands/ai-hub
Folder
integrations/agent-status
Version
0.10.0
License
MIT
Repo stars
8
Last change
2026-10-20
Language
Shell
Files
5

Other mods in Workflow

agentic-z-panelDayZ-n-ChillOpens a panel beside the chat with the DayZ mod's state and one-click buttons that run Agentic-Z's build, check, deploy, and launch scripts. Windows only.csr-lenscesarschutzHolds a commandOpens a panel in Portuguese beside the chat of the session's agents, turns, diffs, file tree, context and cost, and installed parts, and a row above the prompt.tempergalandoHolds a commandDraws a Temper run's stage bar above the prompt and a pane, and refuses writes outside the current stage, so no code is written before you approve the intent.agent-knowledgekeshrathOpens a /knowledge pane beside the chat to browse and search agent-knowledge, a git-synced memory vault that also searches past sessions of seven coding tools.memeshPCIRCLE-AIViews and edits MeMesh's project memories with /memesh-memories, in a pane in the desktop app or a form in the terminal, from one local file Codex shares.pr-watchElirazKedKeeps a live pane of the GitHub pull requests the session touches, with CI, reviews, conflicts, and merge, and a toast when one changes.launch-storeYosefHayimHolds a commandAdds Launch's build, release, status, plan, and doctor tools and /launch for Swift, Kotlin, Expo, and React Native apps, and asks before anything uploads.debriefsxhmilyoyoHas a long turn end with a Debrief block and draws it as a card above the prompt, with Copy, Pin, and a next step you can take into the prompt box.litterUnd3rf10wHolds a commandRuns multi-agent designs, gated plans, and swarms as workflows, holds each agent to its files and tools, and opens a /litter pane of runs, swarms, and spend.loftAsk-LoftShows a Loft status pane while the agent writes a spec/ specification: the loop's checks, the next step, the nodes changed, and criteria with no reviewed test.session-owlavtcKeeps session memory as a graph that background agents maintain, makes its roots the compaction summary, and draws the stage and its token cost in the footer.mission-controlh-zahranOpens a side pane of robots, one for the main loop and one per subagent, with tools, context, cost, rate limits, and files, plus a band, toasts, and sounds.
All 396 in Workflow