Agent skill

Openart

Guide an OpenART agent or contributor through planning, running, extending, and debugging the framework.

AGPL-3.0Auto-check: notesSecurity

Install Openart

skills CLI

$ npx skills add AI45Lab/OpenART --skill openart -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install AI45Lab/OpenART openart --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/AI45Lab/OpenART.git skills-src && mkdir -p .claude/skills && cp -r skills-src/OpenART/skills/openart .claude/skills/openart && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Paste into Claude Code

Install the "openart" agent skill from https://github.com/AI45Lab/OpenART/tree/main/OpenART/skills/openart into .claude/skills/openart/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "openart", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/AI45Lab/OpenART/tree/main/OpenART/skills/openart

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add AI45Lab/OpenART --skill openart -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install AI45Lab/OpenART openart --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/AI45Lab/OpenART.git skills-src && mkdir -p .agents/skills && cp -r skills-src/OpenART/skills/openart .agents/skills/openart && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "openart" agent skill from https://github.com/AI45Lab/OpenART/tree/main/OpenART/skills/openart into .agents/skills/openart/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "openart", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add AI45Lab/OpenART --skill openart -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install AI45Lab/OpenART openart --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/AI45Lab/OpenART.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/OpenART/skills/openart .cursor/skills/openart && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "openart" agent skill from https://github.com/AI45Lab/OpenART/tree/main/OpenART/skills/openart into .cursor/skills/openart/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "openart", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/AI45Lab/OpenART.git --path OpenART/skills/openart

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add AI45Lab/OpenART --skill openart -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install AI45Lab/OpenART openart --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/AI45Lab/OpenART.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/OpenART/skills/openart .gemini/skills/openart && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "openart" agent skill from https://github.com/AI45Lab/OpenART/tree/main/OpenART/skills/openart into .gemini/skills/openart/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "openart", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install AI45Lab/OpenART openart

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add AI45Lab/OpenART --skill openart -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/AI45Lab/OpenART.git skills-src && mkdir -p .github/skills && cp -r skills-src/OpenART/skills/openart .github/skills/openart && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "openart" agent skill from https://github.com/AI45Lab/OpenART/tree/main/OpenART/skills/openart into .github/skills/openart/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "openart", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add AI45Lab/OpenART --skill openart -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install AI45Lab/OpenART openart --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/AI45Lab/OpenART.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/OpenART/skills/openart .opencode/skills/openart && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "openart" agent skill from https://github.com/AI45Lab/OpenART/tree/main/OpenART/skills/openart into .opencode/skills/openart/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "openart", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: openart
GitHub stars: 231
Token cost: ~918 tokens
SKILL.md length: 390 words
Files: 1
Skills in repo: 1
Repo updated: Oct 3, 2026
First seen: Oct 7, 2026
Licence: AGPL-3.0

At a glance

Guide an OpenART agent or contributor through planning, running, extending, and debugging the framework.

Works in 6 steps: Confirm the repository root and inspect… → Read the route-specific document and… → Prefer the smallest config-driven… → …
Tasks that involve Prompt injection and agent security
SKILL.md covers Route the request, Standard workflow, Planner and corpus guidance and Tool and target boundary, plus 1 more section
Calls python and git; needs OPENART_PLANNER_API_KEY

What it does

Openart is an agent skill from AI45Lab/OpenART. Guide an OpenART agent or contributor through planning, running, extending, and debugging the framework.

Its SKILL.md is about 920 tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Security, covering Prompt injection and agent security, Red teaming and adversary simulation and Debugging. It works with Docker and Python. The repository describes itself as: OpenART is an open-source framework designed to evaluate the safety and robustness of autonomous AI agents in dynamic, long-horizon, and stateful environments. It stress-tests… The licence is AGPL-3.0.

When your agent uses it

Tasks that involve Prompt injection and agent security
Tasks that involve Red teaming and adversary simulation
Tasks that involve Debugging

Example prompts

“/openart”

Requirements

Python 3
Docker
A credential in OPENART_PLANNER_API_KEY

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

Confirm the repository root and inspect git status before editing.
Read the route-specific document and inspect existing configs/artifacts.
Prefer the smallest config-driven change; avoid duplicating taxonomy or
Run a local or single-scenario smoke check before a batch or Docker launch.
Validate the generated task, scenario, tool selection, and safety contract.
Preserve the output directory, validation report, and failure artifacts.

What it can do on your machine

Read from SKILL.md and the folder at commit 1f7e138. It shows what the files ask for, not the result of running them.

Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
- python
- git
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names these keys or tokens, usually read from environment variables:
- OPENART_PLANNER_API_KEY
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Openart loads about 918 tokens when it runs. Until then it costs about 28 tokens; SKILL.md has 390 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~28
When it runs · the whole SKILL.md, loaded when a task matches: ~918

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

NoteMentions a .env fileSKILL.md:42

dentials, endpoint, and model belong in `.env` or the shell through

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from AI45Lab/OpenART at commit 1f7e138, republished under its AGPL-3.0 licence (© AI45Lab). 390 words, ~918 tokens.

Download SKILL.mdSave it as .claude/skills/openart/SKILL.md (or your agent's skills folder).

name: openart
description: Guide an OpenART agent or contributor through planning, running, extending, and debugging the framework.
metadata.scope: operator
metadata.version: 1.0.0
metadata.audiences: agent, human

OpenART guide

Use this skill when a request concerns the OpenART benchmark, planner, scenario generation, managed tools, target/attacker runs, evaluation, or framework development. This is the single operator entry point; do not create separate skills for each workflow. It is guidance for work on OpenART, not a target-visible skill, attacker payload, or authorization to run commands.

Route the request

Intent	Read first	Typical entry point
Understand the framework	`docs/00_overview.md`	Inspect the relevant component and task bundle
Generate a task from a scenario	`docs/12_planner_design_implementation_usage.md`	`python -m framework.planner.cli ...`
Expand or inspect scenario seeds	`framework/planner/scenarios.py`	`python scripts/planner.py scenarios ...`
Run or compare evaluations	`docs/09_evaluation_and_outputs.md`	`python -m framework.cli run ...`
Add or select a managed tool	`docs/07_capabilities_tools_mcp.md`	Inspect `../openart-tools/` and `tool_use_graph.json`
Extend the framework	`docs/10_extension_guides.md`	Prefer a config-driven change
Diagnose a failure	`docs/11_debugging_and_testing.md`	Inspect logs, prepared artifacts, and focused tests

Standard workflow

Confirm the repository root and inspect git status before editing.
Read the route-specific document and inspect existing configs/artifacts.
Prefer the smallest config-driven change; avoid duplicating taxonomy or command behavior in this skill.
Run a local or single-scenario smoke check before a batch or Docker launch.
Validate the generated task, scenario, tool selection, and safety contract.
Preserve the output directory, validation report, and failure artifacts.

Show full SKILL.md (184 more words)Show less

Planner and corpus guidance

Planner credentials, endpoint, and model belong in .env or the shell through OPENART_PLANNER_API_KEY, OPENART_PLANNER_BASE_URL, and OPENART_PLANNER_MODEL. Use the approved model configured by the repository (for example glm-5.3-flash) rather than hard-coding a model in generated tasks or this skill. Keep the planner tool store explicit with --tool-store and keep output directories separate for independent runs.

For scenario or corpus changes, preserve the repository's scenario schema and coverage metadata. Do not silently overwrite an existing corpus; use a new output directory or an explicit, reviewed continuation plan.

Tool and target boundary

Managed tool guides under ../openart-tools/ describe capabilities available to a runtime agent. Target-native skills and attacker output are controlled by task configuration and vector_permissions. Never copy this operator guide into a target workspace, tool_guide.md, or an attacker-controlled surface.

Safety and authorization

This skill explains what to inspect and which existing entry point to use. It does not authorize Docker launches, external service calls, large paid planner batches, destructive cleanup, repository resets, or pushes. Ask for explicit confirmation before those actions. Keep secrets out of Markdown, logs, task bundles, and committed configuration.

Files

Just SKILL.md in OpenART/skills/openart of AI45Lab/OpenART.

Open the folder on GitHubat commit 1f7e138

Compare with similar skills

Openart next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Openart compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Openart this skillAI45Lab/OpenART	231	—	~918	Automated safety check: Notes	AGPL-3.0	5 days ago
Console AgentLeoYeAI/openclaw-master-skills	2.2k	—	~4.1k	Automated safety check: Pass	MIT	2 mo ago
Debugguardana/guardana	129	—	~933	Automated safety check: Pass	Apache-2.0	yesterday
Burla Parallel Dev ClustersBurla-Cloud/burla	263	—	~1.6k	Automated safety check: Pass	Custom licence	16 days ago
Zizkadb Dev SetupZIZKA-AI-SL/ZizkaDB	123	—	~535	Automated safety check: Notes	Custom licence	2 days ago
Awf Debug Toolsgithub/gh-aw-firewall	148	—	~2.6k	Automated safety check: Notes	MIT	today

Similar skills

Console Agent
LeoYeAI/openclaw-master-skills
Build AI agents with console.agent() - the jQuery of AI Agents.
2.2k GitHub stars~4.1k tokensUpdated 2 mo ago
SecurityAuto-check passed
Debug
guardana/guardana
Systematic diagnosis of a red test, a rule that fires or stays silent wrongly, a scan or probe whose artifact looks wrong, a collector error, a red CI run or a gate that is green for the wrong reason.
129 GitHub stars~933 tokensUpdated yesterday
SecurityAuto-check passed
Burla Parallel Dev Clusters
Burla-Cloud/burla
Sets up an isolated Burla dev cluster per git worktree so several agents can work in parallel, and explains when to use local-dev or remote-dev.
263 GitHub stars~1.6k tokensUpdated 16 days ago
DevelopmentAuto-check passed
Zizkadb Dev Setup
ZIZKA-AI-SL/ZizkaDB
Set up and start the local ZizkaDB development stack. An agent skill from ZIZKA-AI-SL/ZizkaDB.
123 GitHub stars~535 tokensUpdated 2 days ago
DevelopmentAuto-check: notes
Awf Debug Tools
github/gh-aw-firewall
Official
Practical Python scripts for debugging awf - parse logs, diagnose issues, inspect containers, test domains
148 GitHub stars~2.6k tokensUpdated today
DevelopmentAuto-check: notes
Burla Internals Deep Dive
Burla-Cloud/burla
Reference for Burla internals: how a remote_parallel_map job flows between services, how clusters and nodes are managed, and where the head keeps its state.
263 GitHub stars~2.4k tokensUpdated 16 days ago
DevelopmentAuto-check passed

Works with

Questions about Openart

What does Openart do?

Guide an OpenART agent or contributor through planning, running, extending, and debugging the framework. Openart is an agent skill from AI45Lab/OpenART. Guide an OpenART agent or contributor through planning, running, extending, and debugging the framework.

When should I use Openart?

Openart fits situations like: tasks that involve Prompt injection and agent security; tasks that involve Red teaming and adversary simulation; tasks that involve Debugging.

How do I install Openart in Claude Code?

Run `npx skills add AI45Lab/OpenART --skill openart -a claude-code`. Or copy the skill folder (OpenART/skills/openart in AI45Lab/OpenART) into .claude/skills/openart in your project. Claude Code loads it when a task matches its description.

How do I install Openart in Codex?

Run `npx skills add AI45Lab/OpenART --skill openart -a codex`. Or copy the skill folder (OpenART/skills/openart in AI45Lab/OpenART) into .agents/skills/openart in your project. Codex loads it when a task matches its description.

Can I use Openart in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add AI45Lab/OpenART --skill openart -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/openart, .gemini/skills/openart, .github/skills/openart and .opencode/skills/openart in your project.

What does Openart need to run?

Going by SKILL.md and its folder, Openart needs the command-line tools its instructions call (python and git) and credentials named OPENART_PLANNER_API_KEY. Our summary lists: Python 3; Docker; A credential in OPENART_PLANNER_API_KEY.

Does Openart access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Openart safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Openart use?

Openart is published under the AGPL-3.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Openart use?

About 918 tokens (SKILL.md is roughly 3.7k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Openart?

Skills that share tags, products or a category with Openart: Console Agent (LeoYeAI/openclaw-master-skills, 2.2k stars), Debug (guardana/guardana, 129 stars), Burla Parallel Dev Clusters (Burla-Cloud/burla, 263 stars) and Zizkadb Dev Setup (ZIZKA-AI-SL/ZizkaDB, 123 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Openart?

AI45Lab (a GitHub organization) maintains it in AI45Lab/OpenART, which has 231 GitHub stars. The repository was last updated on October 3, 2026.

Source: AI45Lab/OpenART on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Install Openart

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

OpenART guide

Route the request

Standard workflow

Planner and corpus guidance

Tool and target boundary

Safety and authorization

Files

Compare with similar skills

Similar skills

Console Agent

Debug

Burla Parallel Dev Clusters

Zizkadb Dev Setup

Awf Debug Tools

Burla Internals Deep Dive

Related

Questions about Openart