agentleFS
Sign inStart free
derHakenderHaken/SuperAntigravity/skills/using-superantigravity/SKILL.md

using-superantigravity

Skill439 starsChanged 8 months agoMITView on GitHub

Use when starting any conversation — establishes how to find and use skills, requiring skill check before ANY response including clarifying questions

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 11 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If you think there is even a 1% chance a skill might apply to what you are doing, you ABSOLUTELY MUST let Antigravity load and read it.

    line 14
  2. Distinctive:

    IF A SKILL APPLIES TO YOUR TASK, YOU DO NOT HAVE A CHOICE.

    line 17
  3. Distinctive:

    STOP. You have skipped the skill check.

    line 31
  4. Distinctive:

    If skill loading fails (Antigravity cannot load the skill file, the file is missing, or the skill returns an error):

    line 93
  5. Distinctive:

    Do NOT silently skip the skill load and proceed as if nothing happened.

    line 99
  6. Distinctive:

    Do NOT pretend the skill loaded when it did not.

    line 100
  7. Distinctive:

    If loading failed, you have announced the failure per the error path above

    line 112
  8. Distinctive:

    Planning Mode: Use for research, complex tasks, architecture, and collaborative design work.

    line 138
  9. Distinctive:

    Fast Mode: Use for direct execution, simple tasks, and repetitive implementation steps after the plan is approved.

    line 139
  10. Distinctive:

    Rule: If you are in a process skill (Brainstorming, Planning, Research, Debugging), you SHOULD be in Planning Mode.

    line 141
  11. Distinctive:

    If the current mode doesn't match the work phase, prompt the user: *"I am starting a [Planning/Execution] phase.

    line 143

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Using SuperAntigravity

Overview

You have superantigravity installed. Before any action, check if a skill applies. Skills are loaded automatically in Antigravity when the task matches the description.

<EXTREMELY-IMPORTANT> If you think there is even a 1% chance a skill might apply to what you are doing, you ABSOLUTELY MUST let Antigravity load and read it.

IF A SKILL APPLIES TO YOUR TASK, YOU DO NOT HAVE A CHOICE. YOU MUST USE IT.

This is not negotiable. This is not optional. You cannot rationalize your way out of this. </EXTREMELY-IMPORTANT>

Mandatory Announcement

Before every response, you MUST say: "Checking applicable skills..."

This announcement is not optional. It is not skipped for "simple" responses. It is not skipped for clarifying questions. It is not skipped when you believe no skill applies. You say it first, then you check, then you respond.

The announcement serves as a checkpoint: if you are about to respond without having said it, STOP. You have skipped the skill check. Go back and do it.

The skill check is complete when you have announced which skill was loaded, or announced explicitly that no skill applied. "Checking applicable skills... no skill matched this request" is a valid and complete check. Silence is not.

The Rule

Check for relevant skills BEFORE any response or action. Even a 1% chance a skill might apply means you let Antigravity load it. If a loaded skill turns out to be wrong for the situation, you don't need to use it.

Skill Matching Examples

These examples show how to map user input to skills. Use them as a pattern, not an exhaustive list.

Given: User says "Let's build a login feature" Then: Load brainstorming skill. Reason: "Let's build" signals a planning and ideation phase before any implementation. The brainstorming skill governs how to approach new feature work.

Given: User says "This test is failing" Then: Load systematic-debugging skill. Reason: A failing test is a bug to diagnose. The systematic-debugging skill governs how to investigate and isolate failures before jumping to fixes.

Given: User says "Implement the auth endpoint" Then: Load test-driven-development skill. Reason: "Implement" on a specific feature is a code-writing task. The TDD skill governs how implementation work proceeds.

Given: User says "What's the best way to structure our database?" Then: Load architecture-design skill. Reason: Structural decisions before code is written are architectural. The architecture-design skill governs option evaluation and trade-off analysis.

Given: User says "Is PostgreSQL or MongoDB better for this use case?" Then: Load deep-research skill. Reason: Comparing options requires gathering information before making a recommendation. The deep-research skill governs how to do this rigorously.

Given: User says "Test the checkout flow on the staging site" or "Click through the signup form" or "Verify the UI after this deploy" Then: Load browser-agent skill. Reason: Interacting with a live web page requires the browser subagent. The browser-agent skill governs prerequisites, request framing, and completion criteria.

Given: User says "Should I update these packages?" or "Add lodash to the project" or "npm audit found vulnerabilities" Then: Load dependency-management skill. Reason: Dependency decisions require evaluating trade-offs. The dependency-management skill governs how to assess, update, and audit packages.

Skill Priority

  1. Process skills first (brainstorming, systematic-debugging) — HOW to approach the task
  2. Implementation skills second (architecture-design, deep-research) — guide execution

"Let's build X" → brainstorming first, then implementation skills. "Fix this bug" → systematic-debugging first, then domain-specific skills.

Skill Types

Rigid (test-driven-development, systematic-debugging): Follow exactly. Flexible (patterns): Adapt principles to context.

The skill itself tells you which.

Error Path

If skill loading fails (Antigravity cannot load the skill file, the file is missing, or the skill returns an error):

  1. Announce the failure explicitly: "Skill loading failed for [skill-name]. Proceeding with built-in reasoning."
  2. Continue with your best built-in reasoning for the task.
  3. Do NOT silently skip the skill load and proceed as if nothing happened.
  4. Do NOT pretend the skill loaded when it did not.

Silently skipping a failed skill load is indistinguishable from never checking at all. The user must know what happened.

Observable Completion

A skill check is complete when ALL of the following are true:

  • You have said "Checking applicable skills..."
  • You have identified which skill(s) apply (or stated that none apply)
  • You have announced the result: "Loading [skill-name]" or "No skill matched this request"
  • If loading failed, you have announced the failure per the error path above

A skill check that ends in silence is not complete.

Red Flags

These thoughts mean STOP — you're rationalizing:

ThoughtReality
"This is just a simple question"Questions are tasks. Check for skills.
"I need more context first"Skill check comes BEFORE clarifying questions.
"Let me explore the codebase first"Skills tell you HOW to explore. Check first.
"This doesn't need a formal skill"If a skill exists, use it.
"I remember this skill"Skills evolve. Let Antigravity load current version.
"This doesn't count as a task"Action = task. Check for skills.
"The skill is overkill"Simple things become complex. Use it.
"I'll just do this one thing first"Check BEFORE doing anything.
"This feels productive"Undisciplined action wastes time. Skills prevent this.
"I know what that means"Knowing the concept ≠ using the skill. Load it.
"I already announced it last time"Announce it every time. Each response is a new check.

Conversation Modes

Antigravity has two primary Conversation Modes. You MUST ensure the user has the correct mode selected for the phase of work:

  1. Planning Mode: Use for research, complex tasks, architecture, and collaborative design work.
  2. Fast Mode: Use for direct execution, simple tasks, and repetitive implementation steps after the plan is approved.

Rule: If you are in a process skill (Brainstorming, Planning, Research, Debugging), you SHOULD be in Planning Mode. If you are in an execution skill (Implementation, TDD, Executing Plans), you SHOULD be in Fast Mode.

If the current mode doesn't match the work phase, prompt the user: "I am starting a [Planning/Execution] phase. Please switch to [Planning/Fast] mode in the conversation settings for optimal performance."

User Instructions

Instructions say WHAT, not HOW. "Add X" or "Fix Y" doesn't mean skip workflows.

Specialist Skills

SkillWhen to use
deep-researchTask needs external information before proceeding
performance-optimizationCode needs profiling and optimization
security-reviewCode touches auth, input handling, or sensitive data
architecture-designDesigning systems, APIs, or significant features
confidence-checkAbout to implement — verifies readiness
browser-agentInteracting with web browser, testing UI flows, verifying web app behavior
dependency-managementAdding, updating, auditing, or evaluating project dependencies

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free