A professional-grade negotiation framework for AI agents to represent principals with strict safety guardrails and graduated autonomy levels.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install negotiate
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 negotiate 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 Negotiation Skill?
The Negotiation skill is a sophisticated protocol designed to transform an AI agent into a proxy negotiator. It operates on the core principle of Never Assume, Always Confirm, ensuring the agent never oversteps its bounds or makes unauthorized commitments. By integrating this skill within the ecosystem of Openclaw Skills, users can delegate complex bargaining tasks—ranging from domain acquisitions to salary discussions—while maintaining total control over hard limits, target prices, and walk-away thresholds.
This skill is built for professional and personal use cases where precision and leverage are critical. It prevents common AI pitfalls like revealing budget ceilings or accepting first offers prematurely. Instead, it utilizes a graduated trust model that allows users to scale the agent's responsibilities from a simple observer to an autonomous closer based on established performance and explicit permission.
Negotiation Skill Use Cases
- Managing procurement for high-value assets like domains, NFTs, or specialty hardware.
- Handling peer-to-peer marketplace interactions on platforms such as eBay or Facebook Marketplace to filter lowballers.
- Drafting professional salary and contract counteroffers based on specific BATNA (Best Alternative to a Negotiated Agreement) benchmarks.
- Automating the selling process for products and services while preventing scope creep and unauthorized discounts.
How Negotiation Skill Works
- The user defines a mandatory setup configuration including hard limits, target prices, and walk-away thresholds.
- A specific Category Context is loaded (e.g., buying, selling, or professional) to apply the correct strategic rhythm.
- The agent is assigned an Autonomy Level, defaulting to Level 1 (Observer) where every message must be drafted and approved.
- During interaction, the agent monitors for manipulation tactics like artificial urgency or emotional pressure from the counterparty.
- The agent logs every offer and counteroffer with timestamps, ensuring a transparent audit trail of the negotiation lifecycle.
- Final commitments are only made once the principal provides explicit approval or if the agent has reached Level 4 autonomy within that specific category.
Negotiation Skill Setup
To activate this skill within your Openclaw Skills environment, ensure the core negotiation logic is initialized with the required parameters:
# Example initialization with required parameters
openclaw install negotiate
# Configure your hard limits and target price
openclaw config negotiate --target 350 --limit 500 --category "domain"
Users must explicitly grant autonomy levels per category before the agent can move beyond drafting responses.
Negotiation Skill Data Schema & Taxonomy
The Negotiation skill organizes data through a structured profile and session logs to maintain leverage and historical context.
| Data Component | Description | Example Keys |
|---|---|---|
| Parameters | Core financial and logical boundaries | Hard limit, Target price, Walk-away threshold |
| Autonomy Grants | Record of permissions granted per category | Level 1 (Observer) to Level 4 (Closer) |
| Negotiation Profile | Long-term insights gathered over multiple sessions | Known Limits, Patterns Observed, Past Outcomes |
| Interaction Logs | Detailed record of the current negotiation | Timestamped offers, Counterparty tactics detected |
Negotiation Skill Advanced Features
- Graduated Autonomy: Scale the agent from a passive observer to an autonomous closer across four distinct trust levels.
- Manipulation Detection: Automatically flags tactics such as 'final offer' games and artificial urgency to protect the user's leverage.
- Strategy Defaults: Specialized logic for different markets, including buying.md, selling.md, and professional.md templates.
- Longitudinal Pattern Recognition: Tracks counterparty behavior over time, such as identifying if a buyer typically accepts deals faster on specific days.
SKILL.md
Loading
Related Openclaw Skills
NEET Exam Preparation Assistant
A specialized AI agent skill designed to streamline NEET preparation through data-driven tracking, NCERT-focused scheduling, and mock exam analysis.
NAS Configuration & Management
A technical framework for configuring, securing, and optimizing Network Attached Storage systems with a focus on data resilience and secure remote access.
MySQL
A comprehensive guide for building and maintaining high-performance MySQL databases using production-ready patterns and advanced optimization techniques.
MVP Launch & Validation
A framework for building, validating, and launching lean minimum viable products with strict scope discipline and role-specific guidance.
Neo4j
An expert-level skill for writing performant Cypher queries, managing graph indexes, and executing safe batch operations in Neo4j.
NestJS Expert Guide
A specialized AI coding skill designed to help developers build robust NestJS applications by avoiding architectural traps like circular dependencies and improper DI scoping.
Featured*
