dashbash
DashBash: a live plan dashboard skill for Claude Code.
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Badge for your README
Markdown
[](/listings/niypoo--dashbash)HTML
<a href="/listings/niypoo--dashbash"><img src="/api/listings/niypoo--dashbash/badge.svg" alt="findsafeskills safety rating" /></a>Install
Install instructions from the README
As written by the repo's author. findsafeskills has not verified these steps.
In Claude Code:
/plugin install dashbash --marketplace niypoo/dashbash
On older Claude Code versions, add the marketplace first:
/plugin marketplace add niypoo/dashbash /plugin install dashbash@dashbash
Or clone it as a personal skill:
git clone https://github.com/niypoo/dashbash ~/.claude/skills/dashbash
Start a new session (or run `/reload-plugins`) and it's ready. Python 3 is the only requirement (standard library only). Publishing the live dashboard uses claude.ai Artifacts; without them the page still works as a local HTML file.
Other coding agents (Codex, Gemini CLI, Cursor, Copilot, …)
DashBash's CLI and page don't depend on any model. Clone it anywhere, then teach your agent about it once per project:
git clone https://github.com/niypoo/dashbash ~/dashbash cd your-project python3 ~/dashbash/scripts/dashbash.py agents --for all --agent Codex
That adds a short DashBash section to the files agents read: `AGENTS.md` (Codex, Cursor, Jules, Amp and others), `GEMINI.md` (Gemini CLI), `.github/copilot-instructions.md` (Copilot) and `.cursor/rules/dashbash.mdc` (Cursor). Use `--for agents` (or `gemini`, `copilot`, `cursor`) to write just one. Running it again updates the section in place. `--agent` sets the name the page shows instead of "Claude".
Everything works the same - import, live updates, Progress, Mind map, Timeline, every export - except the claude.ai-only parts: the published link and comments sent straight to the model. On the local page, "Ask or comment" and the task buttons copy a ready message for you to paste into your agent's chat.
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
.claude-plugin/marketplace.json · Manifest · 16 lines
- {
- "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
- "name": "dashbash",
- "description": "DashBash: a live plan dashboard skill for Claude Code.",
- "owner": {
- "name": "niypoo"
- },
- "plugins": [
- {
- "name": "dashbash",
- "source": "./",
- "description": "A live dashboard of your plan that Claude keeps up to date while it works.",
- "version": "1.1.0"
- }
- ]
- }
README.md · README · 108 lines
- # DashBash
- **A live dashboard of your plan that Claude keeps up to date while it works.**
- Ask Claude "where are we?" once. DashBash builds a one-page dashboard of your plan, and from then on Claude updates it every time it finishes something. You always have one link that shows what is done, what Claude is doing now, and what is waiting on you.
- 
- ## Why DashBash
- - **Honest progress.** The percentage counts Claude's work only. Things that wait on you (a decision, a test on your phone, store details, a deploy) get their own card, so they never make the plan look stuck.
- - **Live, not a snapshot.** After every finished task Claude runs one command and republishes the same link. No servers, no extra tabs.
- - **Talk to it.** Every task has an "Ask or comment" button: send a question, a note or an alert straight to the Claude session working on the plan. Tasks that are yours have a check circle: mark one done or report a problem, and Claude updates the plan.
- - **Readable.** Every phase and task carries a one-line description of what it does, so titles never leave you guessing.
- - **Calm by default.** It opens on a short list of sections with one bar each. Click a section for its phases, a phase for its tasks. Two more views: a mind map of the whole plan and a timeline of what got done today, this week, this month or this quarter.
- - **Works with any plan.** Point it at a markdown checklist (`PLAN.md`, `TODO.md`, `ROADMAP.md`) and it reads phases, tasks and real completion dates from your git history. Or let Claude build it from the conversation.
- - **Any language.** English by default; Arabic and right-to-left layouts built in; any other language through a labels file.
- - **Take it with you.** Export a PDF report to send, an Excel workbook with progress bars and colored statuses, a markdown checklist you can import back, a CSV, or the full JSON.
- - **Yours to style.** Light or dark, accent color, eleven Arabic and Latin fonts, text size. Each viewer's choice stays in their browser.
- ## Install
- In Claude Code:
- ```
- /plugin install dashbash --marketplace niypoo/dashbash
- ```
- On older Claude Code versions, add the marketplace first:
- ```
- /plugin marketplace add niypoo/dashbash
- /plugin install dashbash@dashbash
- ```
- Or clone it as a personal skill:
- ```bash
- git clone https://github.com/niypoo/dashbash ~/.claude/skills/dashbash
- ```
- Start a new session (or run `/reload-plugins`) and it's ready. Python 3 is the only requirement (standard library only). Publishing the live dashboard uses claude.ai Artifacts; without them the page still works as a local HTML file.
- ### Other coding agents (Codex, Gemini CLI, Cursor, Copilot, …)
- DashBash's CLI and page don't depend on any model. Clone it anywhere, then teach your agent about it once per project:
- ```bash
- git clone https://github.com/niypoo/dashbash ~/dashbash
- cd your-project
- python3 ~/dashbash/scripts/dashbash.py agents --for all --agent Codex
- ```
- That adds a short DashBash section to the files agents read: `AGENTS.md` (Codex, Cursor, Jules, Amp and others), `GEMINI.md` (Gemini CLI), `.github/copilot-instructions.md` (Copilot) and `.cursor/rules/dashbash.mdc` (Cursor). Use `--for agents` (or `gemini`, `copilot`, `cursor`) to write just one. Running it again updates the section in place. `--agent` sets the name the page shows instead of "Claude".
- Everything works the same - import, live updates, Progress, Mind map, Timeline, every export - except the claude.ai-only parts: the published link and comments sent straight to the model. On the local page, "Ask or comment" and the task buttons copy a ready message for you to paste into your agent's chat.
- ## Use
- In any project, ask Claude:
- ```
- /dashbash
- ```
- or "where are we?", "what's left?", "what's waiting on me?". For another language, name it: `/dashbash ar`, "dashboard in Arabic".
- To stop live updates: "stop the dashboard".
- ### Your plan as a checklist
- DashBash reads plans like this (see [examples/PLAN.md](examples/PLAN.md)):
- ```markdown
- # My App
- ## Foundation
- ### Phase 0: Skeleton
- - [x] T1 Project scaffold
- - [ ] T2 CI pipeline
- ## Release
- ### Phase 1: Store
- - [ ] T3 Store listing
- - [ ] T4 Test on a real iPhone [you]
- ```
- `##` headings become sections, `###` headings phases, checkboxes tasks, and `[you]` marks a task as yours.
- ### The CLI
- Claude drives it, but you can too:
- ```bash
- PD="python3 ~/.claude/skills/dashbash/scripts/dashbash.py" # or the plugin's install folder
- $PD import PLAN.md # build or refresh from a markdown plan
- $PD set T2 done # mark done (stamps the date) and re-render
- $PD summary # 60% of Claude's tasks (3/5) · phases 1/3 · next: ...
- $PD validate # catch missing phases, duplicate ids, missing dates
- ```
- Everything lives in `.claude/status/` in your project: `status.json` (the data) and `dashboard.html` (the page). Add it to `.gitignore` if you don't want it committed.
- Full schema and commands: [reference.md](reference.md).
- ## License
- MIT
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.