Box Ascii logo

Box Ascii

CommunityPopular
davidondrej
box-ascii

Operate Box by Ascii cloud VMs through its REST API, CLI, and SSH. Use for Box provisioning, environments, secrets, templates, snapshots, stop/resume/fork, or running BB and coding agents on box.ascii.dev. This is for Ascii's compute service, not Box.com file storage.

Overview

Publisherdavidondrej
Repositoryskills
Skill namebox-ascii
Stars
4.1K
Forks
599
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by davidondrej on GitHub. Read the source before you install it.

Installation

Install the Box Ascii AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/davidondrej/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/skills/ops-and-setup/box-ascii .claude/skills/box-ascii
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Box Ascii in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Box Ascii on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Box Ascii is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Box by Ascii

Use Box's API or CLI for machine management and SSH or the command API for work inside a VM. Read operations and observations before direct API calls, credential changes, or recovery work.

Start with the actual account and machine

  1. Read the current project's instructions and operational notes. Identify the intended account, organization, Box, and repository; retain existing mappings instead of provisioning duplicates.
  2. Load BOX_API_KEY from the environment or the project's documented, gitignored credentials file. Read its location from the project's setup instructions; do not assume a universal path. Do not search unrelated credential stores or print their contents. Use the secrets skill if a new key is needed.
  3. For operations, check authentication and state with GET /me, /limits, and /boxes. If the CLI is available, inspect box status, box list, and the relevant --help instead. Filter output to the fields needed for the task.
  4. Check which tools are already installed. An API workflow does not require the Box CLI or an SDK. Read the official install instructions if installation is needed; do not execute a remote installer blindly.

Use the DeepAPI skill to read current Box docs when behavior or request fields are uncertain. Prices, limits, model catalogs, and install commands must be checked live. A documentation lookup is not evidence that an authenticated operation succeeded.

API access

  • Base URL: https://ascii.dev/api/box/v1.
  • Authentication: Authorization: Bearer $BOX_API_KEY.
  • JSON bodies: Content-Type: application/json.
  • Use an account API key for account-wide operations. A machine-scoped key is restricted to its Box.
  • Key creation, rotation, and revocation require a browser sign-in session, even through box api-key. An existing API key cannot create another key.
  • Do not assume provider onboarding has an API. The documented Box-managed provider login flow uses the Agents page; injecting already configured agent credentials is a separate environment setting.

This read-only example checks authentication without printing account details or credentials. Load the key into the process environment first:

python
import json
import os
import urllib.error
import urllib.request

request = urllib.request.Request(
    "https://ascii.dev/api/box/v1/me",
    headers={"Authorization": "Bearer " + os.environ["BOX_API_KEY"]},
)
try:
    with urllib.request.urlopen(request, timeout=30) as response:
        data = json.load(response)
        print({"status": response.status, "ok": data.get("ok")})
except urllib.error.HTTPError as error:
    print({"status": error.code})
    raise SystemExit(1)

Python's standard library worked in our tests. Use the official TypeScript or Python SDK when the surrounding application benefits from typed clients; do not introduce a client abstraction just to make a few requests.

Preserve the user's scope

  • Write a focused assignment preserving intent, constraints, and corrections. Never invent requirements, deliverables, permissions, or approval.
  • Model, machine, repository, and thread placement guide the launcher. Give each worker only its task and necessary context.
  • Before sending, remove copied conversation, launcher instructions, and unrelated assignments. Never prepend the full user prompt; quote exact wording only when explicitly requested.

Environments and templates

An environment holds repositories, secrets, and credential permissions. A named snapshot/template holds installed tools and disk state. Use both for repeatable startup.

Common CLI operations, after checking the target and current contents:

bash
box env list
box env new dev
box env add-repo dev owner/repository --branch main
box env set-file dev repository/.env --from ./private/project.env
box env set dev --github true --secrets true --box-credentials false --agents-credentials true
box new --environment dev
  • Repository branches are cloned as configured; Box does not create a working branch. Let the coding workflow or BB manage branches/worktrees.
  • Repositories land at /home/user/<repository-name>. Secret file paths are relative to /home/user; include the repository folder to place a .env inside its clone.
  • --environment dev selects the environment. --env KEY=value sets a variable for one Box; it overrides the environment value. Never put real secrets in CLI arguments.
  • Named environments must already exist. Omitted selection uses the default for a new Box/template deploy; ordinary resume and fork preserve the existing/source environment version.
  • Saving an environment creates a version. Existing Boxes do not automatically receive it, including on ordinary resume. Inspect box info and upgrade deliberately.
  • box env upgrade dev can affect many Boxes. For one Box, use the API's explicit agentIds selection (the schema calls Box IDs "agent" IDs). Withheld secrets are deleted on upgrade; re-pinning an older version does not restore deleted secret files.
  • Editing an environment also changes what future Boxes receive, even if only one existing Box is upgraded. For an exception intended for one Box alone, use Box-specific configuration instead of changing a shared environment.
  • After injection, restart the consuming application if it reads variables only at process start. Check the running application's behavior, not only a new SSH shell.

Preinstall BB, required runtimes, and dependencies in a clean template. Keep credentials in environments or inject them for the intended user after boot. Set BOX_ID to the verified prepared Box's identifier before these commands.

bash
box snapshot "$BOX_ID" agent-stack
box new --from agent-stack --environment dev

Verify the saved snapshot is ready before deploying it. Re-saving the same name changes future deploys, not existing machines. Fresh repository state and successful dependency installation still need validation.

Isolation and credentials

  • For personal Boxes, inject only the credentials and repos the task needs. For Boxes driven by other users, use an environment marked safeForThirdParties: true or noEnv: true.
  • Protected mode withholds the owner's managed GitHub, secrets, Box, and agent credentials regardless of the section toggles. Explicit per-Box env values still pass through.
  • Do not promise that protected mode sanitizes arbitrary disk contents. Manually added cloud, package-manager, and Docker logins may remain in a template or snapshot.
  • Never build a shared template from a credential-filled personal machine. Per-Box variables also carry into forks/template deploys unless replaced.
  • Conversion to protected mode scrubs managed credential files, including Codex/Claude logins that the Box's own user may have added. Conversion is one-way. Check ownership and recovery needs before applying it to an existing Box.
  • Keep keys, auth files, enrollment codes, desktop URLs, and raw API/session logs out of prompts, committed files, and user-visible output. Desktop URLs can contain access tokens.

Verify the requested outcome

After a change, inspect the specific Box and verify the result inside it. Check file presence/permissions without printing secrets. For agent setup, run a harmless task and verify the intended account, repository, output, and continued conversation.

Report the Box/environment identifier, what changed, what actually passed, and any real limitation. Do not equate an accepted API request, an SSH-ready machine, or a snapshot with a working agent session.

Official references

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Box Ascii AI skill do?

Operate Box by Ascii cloud VMs through its REST API, CLI, and SSH. Use for Box provisioning, environments, secrets, templates, snapshots, stop/resume/fork, or running BB and coding agents on box.ascii.dev. This is for Ascii's compute service, not Box.com file storage.

Why use Box Ascii on TypingMind?

Because you install it once and use it with any model. Box Ascii is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Box Ascii in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/davidondrej/skills/tree/main/skills/ops-and-setup/box-ascii. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Box Ascii?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Box Ascii?

As many as you like. As long as a model supports skills, you can use Box Ascii with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Box Ascii AI skill free?

Yes. It is published on GitHub by davidondrej under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇