EASYCLAW INSTRUCTIONS FOR USE

• --Provider: Easylab AI
• --Contact: jdoussot@easylab.ai
• --Registered address: Luxembourg
• --System: EasyClaw -- Autonomous AI Agent Platform

EasyClaw is a general-purpose autonomous agent platform (AI deployer application, Article 26) intended for:

• --Business process automation
• --Administrative task execution
• --Research and information synthesis
• --Multi-channel communication management
• --IT operations and system administration
• --Internal knowledge management

EasyClaw is NOT intended for:

• --Any purpose prohibited under Article 5 of the AI Act
• --Autonomous decision-making affecting natural persons' legal rights
• --Medical diagnosis or treatment recommendations
• --Financial advice or automated trading
• --Any application listed in Annex III without appropriate safeguards

EasyClaw operates as an autonomous agent capable of executing multi-step tasks through an orchestration layer built on OpenAI GPT-5.4 LLM. The system maintains persistent memory across sessions, communicates over multiple channels, and executes system operations within defined security boundaries.

Performance is continuously monitored via a real-time dashboard tracking uptime, trust score, security incidents, and memory integrity. Historical metrics are retained for audit purposes.

• --Outputs may contain inaccuracies and should be verified by competent personnel
• --System relies on underlying LLM which may exhibit biases present in training data
• --Real-time data access limited to configured tools and integrations
• --Performance may vary based on task complexity and resource availability
• --Not suitable for safety-critical autonomous operations without human oversight

Deployers must:

• a)Designate qualified oversight personnel with appropriate training on AI system operation and limitations
• b)Provide said personnel with access to the EasyClaw monitoring dashboard at all times during system operation
• c)Configure the exec-guardian PIN system for their deployment environment, ensuring PIN holders are identified and accountable
• d)Establish written procedures for system halt scenarios, including criteria for intervention and escalation paths
• e)Review automated operation logs on a regular basis, no less than weekly, with documented findings

• --Text-based input only; the system does not process images, audio, or video
• --No biometric data shall be provided as input under any circumstances
• --Deployer is responsible for ensuring data quality and representativeness of inputs
• --Sensitive personal data should be minimised in accordance with GDPR principles
• --Input data is processed locally; inference calls to the LLM provider are subject to their data processing terms

• --OpenClaw framework updates are provided by Easylab AI and documented in release notes
• --LLM model updates by OpenAI may affect system behaviour; deployers will be informed of material changes
• --Deployers will be notified of any updates that materially change system capabilities or risk profile
• --Local configuration changes are the deployer's responsibility and should be documented
• --Security patches are issued as needed and should be applied promptly

• --Automatic logging of all agent operations is enabled by default and must not be disabled
• --Operation logs are retained for a minimum of six (6) months
• --Dashboard access is provided to designated oversight personnel upon deployment
• --Security alerts are configured for anomalous behaviour patterns, including unusual command execution frequency, failed authentication attempts, and rate limit breaches

• --Compliance inquiries: jdoussot@easylab.ai
• --Technical support: Provided per the terms of the deployment agreement
• --Incident reporting: Via the monitoring dashboard or direct contact with the provider