ワンクリックで
agy-customizations
>-
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
>-
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Begin a new implementation phase. Activates when the user invokes /start-phase or asks momentum to run the start-phase recipe.
Work with concurrent workstreams (Rule 15). Activates when the session involves multiple active branches, cross-lane coordination, or when starting work that should not live on the current branch.
Record a new backlog item (bug, feature, tech debt, or enhancement) when you discover work that needs tracking. Activates when the session identifies unplanned work that should be tracked in the backlog.
Check spec structure health of the momentum project. Activates when the user runs `/validate`, asks for a health check, or when spec changes need verification before phase completion.
Drive a single repo through its phase as part of a swarm. Activates when spawned by the conductor (the user's primary session) with a spawn directive. State lives in files; the conductor reads board.json between turns. Loads on the swarm-supervisor invocation only — not invoked directly by the user.
Read specs/status.md before starting any work in a momentum-managed project. Activates at the start of every new task to ensure the agent has current phase context. Codifies Rule 1 (Orient First).
| type | Evidence |
The Antigravity Customization System allows you to tailor the agent's behavior, teach it new workflows, enforce guidelines, and integrate it with external tools. By customizing the agent, you can transition it from a general-purpose assistant to an expert pair programmer specialized in your project's codebase and processes.
Choose the right customization type based on your goal:
| Type | Config File/Folder | Scope | Best For | Learn More |
|---|---|---|---|---|
| Rules | GEMINI.md, AGENTS.md | Contextual / Hierarchical | Enforcing coding styles, API restrictions, and local guidelines. | Rules Guide |
| Skills | skills/<name>/SKILL.md | On-Demand (Progressive) | Teaching the agent multi-step procedures, runbooks, and tool workflows. | Skills Guide |
| Plugins | plugins/<name>/plugin.json | Bundle | Packaging related skills, rules, and MCP configs into a single unit. | Plugins Guide |
| Hooks | hooks.json | Lifecycle Event | Running scripts/commands at specific agent lifecycle points (e.g., pre-tool execution). | Hooks Guide |
| MCP Servers | mcp_config.json | Tool Integration | Connecting the agent to external services and custom tool providers. | MCP Guide |
Antigravity automatically discovers customizations by traversing specific directories.
.agents/ (or .agent/, _agents/, _agent/) at the root of
your project..git) to find these directories.GEMINI.md, AGENTS.md, .agents/rules/*.md~/.gemini/config/When multiple customizations are discovered, they are loaded and applied in a specific order. If there are naming conflicts (e.g., two skills with the same name), the higher-priority customization overrides the lower-priority one.
The priority order (from highest to lowest) is:
skills.json or plugins.json in your workspace.~/.gemini/config/To prevent overwhelming the model's context window, Antigravity uses progressive disclosure:
trigger: model_decision behave similarly. Only always_on
rules are loaded unconditionally.All customizations (especially rules) are deduplicated by their resolved file paths. A rule file will never be injected more than once in a single conversation turn, even if it matches multiple trigger conditions.
For customizations stored in non-standard locations, you can use skills.json
and plugins.json to explicitly register them and inherit from shared
configurations.