Agent Doppelgänger is a constrained autonomous delegate that manages cross-platform communications while enforcing strict identity fidelity and safety policies.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-doppelganger
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 agent-doppelganger 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 Agent Doppelgänger (ADG)?
Agent Doppelgänger (ADG) acts as a sophisticated, policy-bounded identity proxy designed to handle real-world communication on your behalf. By integrating this capability into your collection of Openclaw Skills, you can delegate interactions across platforms like Email, Discord, Slack, and WhatsApp without losing your personal touch or risking unauthorized actions. ADG operates as a constrained autonomous delegate, meaning it only communicates within formally provable limits defined by your specific declarative policies.
The skill excels at maintaining identity fidelity through a multi-layered modeling approach. It doesn't just generate text; it analyzes intent, evaluates authority, and applies style heuristics to ensure every interaction feels authentic while remaining strictly within the boundaries you set. This makes it an essential tool for developers and professionals looking to scale their digital presence securely.
Agent Doppelgänger (ADG) Use Cases
- Automating responses to routine queries on Slack or Discord while maintaining a consistent professional tone.
- Filtering and responding to incoming emails based on predefined heuristics and decision logic.
- Delegating low-stakes communication tasks across multiple fragmented channels through a single identity proxy.
- Ensuring high-stakes or sensitive requests are automatically escalated for manual human review.
How Agent Doppelgänger (ADG) Works
- Adapter: The system normalizes incoming messages from various channels like WhatsApp or Email into a standardized internal format.
- Intent Analysis: The agent classifies the message intent based on domain, stakes, authority requirements, and ambiguity level.
- Policy Gate: Before any content is generated, ADG evaluates declarative policies via a DSL to determine if it has the authority to handle the request.
- Confidence Engine: A scoring system measures the reliability of the intent analysis and the proposed response strategy.
- Response Generation: The agent synthesizes a message using your unique Style, Heuristics, and Preferences to maintain identity fidelity.
- Verifier: A final audit layer checks the response against hard constraints and forbidden domains before the message is sent or drafted.
Agent Doppelgänger (ADG) Setup
To deploy Agent Doppelgänger as part of your Openclaw Skills workflow, follow these installation steps:
# Clone the ADG repository
git clone https://github.com/openclaw/agent-doppelganger.git
cd agent-doppelganger
# Install dependencies and build the identity model
npm install
npm run build -- --profile=your-identity-config
Ensure you have configured your adapter credentials for specific platforms in the config/ directory before initializing the delegate.
Agent Doppelgänger (ADG) Data Schema & Taxonomy
ADG organizes identity and communication data through a structured four-layer taxonomy:
| Layer | Component | Function |
|---|---|---|
| Style | Surface Characteristics | Manages vocabulary, sentence length, and directness. |
| Heuristics | Decision Logic | Core rules like "never accept meetings without an agenda." |
| Preferences | Soft Weights | Prioritizes certain domains, such as favoring Work over Social requests. |
| Constraints | Hard Rules | Fixed boundaries that trigger immediate escalation if approached. |
All session metadata is logged with intent classifications (Domain, Stakes, Authority) for transparency and auditing.
Agent Doppelgänger (ADG) Advanced Features
- Declarative Policy Language: Uses a custom DSL evaluated prior to generation to prevent prompt injection and policy overreach.
- Automatic Escalation: Sophisticated logic that blocks or drafts messages when confidence is low or forbidden domains (e.g., Legal, Finance) are detected.
- Cross-Platform Normalization: Unified handling of disparate communication protocols including Email, Discord, and Slack.
- Identity Fidelity Modeling: Advanced synthesis of personal heuristics and preferences to ensure the autonomous delegate reflects the user's true persona.
SKILL.md
Loading
Related Openclaw Skills
Location Safety Monitor
A proactive safety monitoring system that tracks user location and host health to provide automated alerts and emergency escalation.
AI Tech Lead & Architect
A sophisticated AI orchestration skill that applies strict Context Engineering and a four-phase methodology to deliver maintainable, secure, and architecturally sound code.
Luna Calorie Tracker
An AI nutrition assistant that uses vision analysis to log calories and macros directly from food photos.
TiDB X
TiDB X is an object-storage-native, elastic, and multi-tenant SQL engine designed specifically for AI agent workloads with query-based billing.
BlackSnow
BlackSnow is an economic sensor engine that converts fragmented data exhaust into machine-readable, tradable risk primitives.
Media Orchestrator
A unified skill for resolving, downloading, and delivering high-quality audio and video content to chat platforms via Openclaw Skills.
Featured*
