- Antigravity is a built-in Easter egg skill that ships with every OpenClaw installation — no setup required
- It responds to a trigger keyword with a playful message, bypassing the LLM entirely — zero token cost, instant response
- The skill's SKILL.md file is the simplest possible example of OpenClaw's skill format — one trigger, one response
- Copying and modifying it is the fastest way to build your first custom skill from scratch
- As of early 2025, Antigravity loads automatically on startup and works in every channel type OpenClaw supports
Antigravity is the skill that makes new OpenClaw builders stop and grin. Trigger it once, see the response, and you immediately understand something fundamental about how this system works. It's a deliberate teaching tool disguised as a joke — and the joke lands every time.
What the Antigravity Skill Actually Is
OpenClaw Antigravity is a lightweight built-in skill that responds to a specific trigger keyword with a humorous, self-aware message. The name is a nod to Python's famous import antigravity Easter egg — type that into any Python interpreter and it opens a webcomic about the joy of programming. OpenClaw's version plays on the same tradition: a tiny, unexpected delight baked into a serious tool.
The skill lives in OpenClaw's built-in skill library, which means it loads automatically every time OpenClaw starts. You don't install it. You don't configure it. It's just there, waiting.
What makes it genuinely interesting isn't the joke itself — it's the mechanism. Antigravity demonstrates that a skill doesn't need an LLM, doesn't need an API call, and doesn't need more than a handful of lines in a SKILL.md file. The gateway matches the trigger keyword directly and returns the configured response. No model round-trip. No latency. Instant.
The Antigravity skill proves a point: not every agent response needs to go through a language model. For commands, shortcuts, static lookups, and canned responses, a simple keyword-triggered skill is faster, cheaper, and more reliable than any LLM call. This is the pattern to reach for whenever you know exactly what the answer should be.
How to Trigger It and What Happens
Triggering Antigravity is exactly what you'd expect. Send a message containing the skill's configured keyword — by default, antigravity — to any agent channel that has the skill loaded. The agent responds immediately with a playful message about defying gravity, productivity, and the nature of AI agents.
Here's what the interaction looks like in a Telegram channel:
User: antigravity
Agent: 🚀 Activating antigravity protocol...
Laws of physics? Suggestions, really.
Your agent is now operating in zero-gravity mode.
(Results may float.)The exact response text is defined in the skill's SKILL.md file and can be changed to anything you want. The trigger keyword is also configurable. Some teams rename the skill entirely and use the same pattern for internal slash commands.
Sound familiar? If you've built command bots in Slack or Discord, this is the exact same pattern — a keyword matcher that fires a canned response. OpenClaw formalizes it into the SKILL.md format so every skill, no matter how simple, follows the same structure as your most complex multi-step agent workflows.
The SKILL.md File Anatomy
Every OpenClaw skill is defined by a SKILL.md file. The Antigravity skill's SKILL.md is the shortest possible version of this format — which is exactly what makes it the best starting point for understanding how skills work.
---
name: Antigravity
trigger: antigravity
description: Easter egg skill. Responds to the antigravity keyword.
llm: false
---
# Antigravity
When triggered, respond with:
🚀 Activating antigravity protocol...
Laws of physics? Suggestions, really.
Your agent is now operating in zero-gravity mode.
(Results may float.)The frontmatter block (between the --- delimiters) contains the metadata OpenClaw reads at startup. The name is what appears in skill lists. The trigger is the keyword the gateway watches for. The description is shown in the skill directory. And llm: false tells OpenClaw to bypass the language model entirely — this skill runs in pure keyword-match mode.
The body below the frontmatter is the skill's instructions. For LLM-powered skills, this section contains the system prompt. For non-LLM skills like Antigravity, it contains the exact response text the agent should return.
The llm: false flag is powerful but narrow. Use it only for skills where the response is always the same regardless of context. The moment your skill needs to interpret user input, generate a dynamic response, or make a decision — set llm: true and write a proper system prompt. Bypassing the model on a skill that needs it produces incorrect, confusing responses.
Practical Workflows Built on the Antigravity Pattern
The Antigravity pattern — keyword trigger, static response, no LLM — is surprisingly useful beyond Easter eggs. Here are three real workflows built on the same mechanism.
1. Internal Command Shortcuts
Teams create "command skills" that respond to short keywords with formatted reference information. Trigger standup and the agent returns the standup meeting template. Trigger deploy-checklist and it returns the deployment checklist. These are things your team knows by heart but still has to look up — a keyword skill puts them one message away in any channel.
2. Status Page Links
Some teams wire a status trigger to return the direct URL for their status page, incident tracking dashboard, and on-call rotation link. No LLM needed. No API call. The response is always the same — just formatted text with links. Instant, free, always accurate.
3. Onboarding Prompts
New team members trigger start or help and receive a structured list of the most important skills, channels, and links in the OpenClaw setup. This works because the onboarding content doesn't change often, and when it does, updating the SKILL.md is a one-minute task.
| Use Case | Trigger | LLM? | Token Cost |
|---|---|---|---|
| Easter egg / fun | antigravity | No | Zero |
| Standup template | standup | No | Zero |
| Deploy checklist | deploy-checklist | No | Zero |
| Onboarding guide | start | No | Zero |
The pattern scales. Once your team knows how SKILL.md files work, anyone can add a new command skill in under five minutes. No code. No deployment pipeline. Drop a file in the skills directory and restart OpenClaw.
Common Mistakes With Simple Skills
- Choosing a trigger keyword that collides with natural language — if your trigger is
help, it fires on any message containing the word "help," including mid-sentence uses. Use specific, unlikely-to-appear-naturally keywords or add a prefix like/help. - Putting reasoning instructions in an llm: false skill — the model is bypassed entirely. If your SKILL.md body contains logic like "if the user asks X, say Y," it will be ignored. The entire body is returned as a static response.
- Forgetting to restart OpenClaw after adding a SKILL.md — skills are loaded at startup. Changes to SKILL.md files don't hot-reload. Restart the process to pick up new or modified skills.
- Duplicating trigger keywords across skills — OpenClaw loads the first matching skill it finds. If two skills share a trigger, the second one never fires. Keep triggers unique across your entire skill directory.
- Removing the Antigravity skill in production without testing — not because it matters functionally, but because every new builder you onboard will try to trigger it and wonder why it doesn't work. Keep it. It costs nothing and builds culture.
Frequently Asked Questions
What does the OpenClaw Antigravity skill actually do?
The Antigravity skill is a playful Easter egg bundled with OpenClaw. When triggered, it responds with a whimsical message referencing the Python antigravity module tradition. It demonstrates how trivially simple a SKILL.md file can be — no API calls, no config, just a pure text response defined in markdown.
How do I trigger the Antigravity skill in OpenClaw?
Send any message containing the trigger keyword — typically antigravity — to any agent channel. Your agent responds immediately. No additional setup beyond having the skill installed is required. It works in every channel type: Telegram, WhatsApp, or the REST API.
Is the Antigravity skill included by default in OpenClaw?
Antigravity ships with OpenClaw's built-in skill library and loads automatically on startup. You can disable it by removing or renaming its SKILL.md file if you want a cleaner skill list in production deployments.
Can I modify the Antigravity skill to create my own Easter egg?
Absolutely. The Antigravity SKILL.md is a perfect template for any simple trigger-response skill. Copy the file, rename it, change the trigger and response text, and restart OpenClaw. Your custom skill loads alongside the original without any conflict.
Does the Antigravity skill use any LLM tokens?
No. Skills with llm: false bypass the language model entirely. The gateway matches the trigger keyword and returns the configured response directly — instant and free. This pattern is worth using for any response that never needs to vary based on context.
What does Antigravity teach about building custom OpenClaw skills?
It shows that a skill can be as minimal as a trigger keyword and a static response. This pattern is the foundation for command shortcuts, quick references, and template responses. Master this first — then layer in LLM calls, API requests, and multi-step logic as your skills grow.
S. Rivera builds and maintains production OpenClaw deployments for teams across fintech, legal, and media. Has authored over 40 custom skills ranging from trivial keyword shortcuts to complex multi-agent orchestration pipelines, and runs internal skill-building workshops for engineering teams adopting OpenClaw.