GClaw for Openclaw

A sophisticated Gmail intelligence tool for Openclaw Skills that enables heuristic email classification and automated digestion without LLM overhead.

yhyatt
v1.0.1
Mar 7, 2026
0
0
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install gclaw

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install gclaw 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 GClaw?

GClaw provides a robust framework for bringing Gmail intelligence into the Openclaw ecosystem. It is designed to read, parse, and categorize email data so that important information surfaces without the noise of a crowded inbox. By focusing on efficiency, GClaw allows developers to build agents that interact with email history and incoming messages through a structured, bot-specific context.

This skill is particularly valuable for developers building Openclaw Skills that require high-volume email processing. It utilizes a heuristic classifier to sort messages into 15 distinct categories—such as finance, travel, and work—ensuring that AI agents only spend LLM tokens on high-value tasks like summarization or narrative generation rather than basic sorting.

GClaw Use Cases

  • Creating automated daily digests that synthesize newsletters and work updates.
  • Building specialized travel bots that monitor specific senders like booking.com or airlines.
  • Developing financial tracking agents that isolate receipts and invoice notifications.
  • Filtering high-traffic inboxes to surface only action-required items for AI coding agents.
  • Powering multi-agent workflows where different bots require isolated views of the same Gmail account.

How GClaw Works

  1. The BotContext defines the specific scope for an agent, including allowed Gmail labels and sender filters.
  2. EmailFetcher utilizes the gog CLI to securely retrieve messages while applying deduplication logic to avoid redundant processing.
  3. EmailParser extracts metadata, cleans the message body, and identifies forwarded content for cleaner data ingestion.
  4. EmailClassifier applies a heuristic engine to assign the email to one of 15 pre-defined categories without incurring LLM costs.
  5. EmailStore persists the processed data into bot-specific JSONL files, making the structured data available for downstream Openclaw Skills.

GClaw Setup

To integrate GClaw into your Openclaw Skills environment, ensure the gog skill is installed and authenticated. Set your environment variables and install the package:

export GCLAW_GMAIL_ACCOUNT="[email protected]"
export GOG_KEYRING_PASSWORD="your-keyring-password"

pip install -e .

You can verify the setup by running the included test suite:

pip install pytest
pytest tests/

GClaw Data Schema & Taxonomy

GClaw maintains a clean data taxonomy to ensure compatibility across different Openclaw Skills. Data is organized as follows:

Component Functionality
BotContext Manages per-bot label isolation and result limits
EmailClassifier Labels data as newsletter, finance, travel, work, etc.
EmailStore Saves deduplicated records in JSONL format
Metadata Includes sender details, cleaned body text, and forward detection

All data is stored with per-bot isolation, allowing multiple agents to query the same account without data leakage or interference.

GClaw Advanced Features

  • Heuristic Classification: Sorts emails into 15 categories including security, health, and legal with zero LLM token cost.
  • BotContext Isolation: Permits multiple bots to share a single Gmail account with unique label and sender restrictions.
  • Deduplication: Built-in persistence logic ensures that no email is processed or summarized more than once.
  • High-Performance Parsing: Advanced body cleaning and forward detection to provide clean text for Openclaw Skills agents.
  • JSONL Persistence: Streamlined data storage for easy integration with external data pipelines and analytics tools.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*