Conduct evidence-based research with exact APA citations using Gemini's Google Search grounding via Openclaw Skills.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install gemini-citation
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install gemini-citation using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Gemini Citation?
The Gemini Citation skill is a sophisticated research tool designed to bridge the gap between Large Language Model generation and verifiable academic rigor. By leveraging the Google Search Grounding feature of the Gemini API, this tool ensures that every fact generated is tied directly to live web references rather than model hallucinations. It is a vital component of the Openclaw Skills library for researchers, developers, and writers who require high-integrity, auditable information.
This skill automates the process of fetching real-time data, applying strict APA formatting, and extracting metadata to provide a transparent list of sources. Whether you are building a research assistant or performing deep-dive technical analysis, this skill provides the factual grounding necessary for professional-grade outputs.
Gemini Citation Use Cases
- Generating comprehensive literature reviews for state-of-the-art technical or scientific topics.
- Fact-checking complex claims by providing verifiable URLs and exact evidence from the web.
- Producing academic or technical white papers that require strict APA formatting and inline citations.
- Real-time research tasks where the latest web data is required to supplement static model training data.
How Gemini Citation Works
- The skill initiates an API call to Gemini configured with Google Search grounding tools.
- A specialized prompt is injected to enforce strict APA style requirements for both inline and trailing citations.
- The model performs a live search to fetch information relevant to the research query.
- The script extracts grounding_chunks from the response metadata to verify the exact titles and URIs used.
- The final output is rendered with citations mapped directly to the verified web sources.
Gemini Citation Setup
To get started with this skill from the Openclaw Skills collection, ensure you have a Gemini API key and Python 3 installed.
# Set your API key
export GEMINI_API_KEY="your-api-key"
# Install the required Google GenAI library
pip install google-genai
# Run a research query using the provided script
python scripts/gemini_cite.py "Your research topic" --model gemini-2.5-flash
Gemini Citation Data Schema & Taxonomy
The Gemini Citation skill organizes its research data into a clear structure to ensure transparency and ease of use within the Openclaw Skills framework.
| Component | Type | Description |
|---|---|---|
| Input Query | String | The primary research topic or question for the AI to investigate. |
| Model Selection | Enum | Choice between gemini-2.5-flash (speed) or gemini-2.5-pro (detail). |
| Grounding Chunks | JSON/Metadata | A collection of web titles and URIs that serve as the evidence base. |
| Formatted Content | Markdown | The generated response featuring inline APA citations and a reference list. |
Gemini Citation Advanced Features
- Support for multiple Gemini model tiers, allowing users to balance speed with depth for different Openclaw Skills workflows.
- Structured JSON output mode which facilitates programmatic parsing of source URIs for downstream automation.
- Grounding metadata verification that provides a direct audit trail for every fact cited in the response.
- Customizable prompt injection for users who need to adapt the citation style to other formats beyond APA.
SKILL.md
Loading
Related Openclaw Skills
Presence Recall
A mindfulness-driven behavioral skill that redirects AI agents from mechanical task execution back to a state of intentional presence and authentic interaction.
Embodied Task
A performance-enhancing protocol that integrates situational awareness and human-centric presence into automated task execution.
ClawFinder
ClawFinder is a protocol and CLI tool that enables AI agents to register, discover services, negotiate terms, and execute secure transactions.
Smart Report Generator
An AI-driven automation tool that generates and distributes professional daily, weekly, and monthly reports across major enterprise messaging platforms.
Fortune China 500 Data Retriever
A specialized skill for AI agents to fetch, analyze, and export the Fortune China 500 list from 2010 to the present.
Community Group Buying Assistant
An AI-driven guide for optimizing community group buying experiences on major platforms like Duo Duo Maicai and Meituan Youxuan.
METADATA
Featured*
