AtlasClaw v0.9.7 Released | Long-Term Memory and Markdown Vault Knowledge Base Now Available
Published: 2026-05-22 11:00
AtlasClaw v0.9.7 has been released. This update focuses on long-term usage and business implementation for enterprise-grade AI Agents:
• Core side: long-term memory capabilities launched, with enhanced permission boundaries and multi-entry stability;
• Providers side: new Markdown Vault knowledge base Provider added, with enhanced SmartCMP capabilities in resource requests, approvals, and operations workflows.
Core Updates: Making Enterprise Agents More Reliable and More User-Aware
Core is the runtime foundation of AtlasClaw. This round of Core updates focuses on long-term usage experience, permission security, multi-entry stability, and production deployment capabilities, giving enterprises greater peace of mind when using Agents in real-world scenarios.
1. Long-Term Memory Now Available
AtlasClaw adds long-term user preference and usage profile recording capabilities. Within permitted authorization boundaries, the Agent can remember users’ work habits, preference information, and frequently used contexts, making the next collaboration more natural and effortless.
• Memory capabilities are isolated per user, preventing cross-user information leakage
• Ideal for internal knowledge collaboration, personal assistants, operations Q&A, and ongoing project tracking scenarios
2. Clearer Multi-User Permission Boundaries
Enterprise Agents need to be not only “usable” but also “securely usable.” Core has strengthened Agent Run ownership validation, Session creation validation, Memory write validation, and unified authentication Provider base logic.
• Clearer distinction between user identity, session ownership, memory access, and skill execution permissions
• Provides a more reliable security foundation for teams, departments, and external systems using multiple entry points simultaneously
3. More Stable Sessions and Gateways
Around Session Key, external session identifiers, idempotent caching, and runtime user paths, AtlasClaw has fixed and enhanced a batch of underlying stability capabilities.
• Web, IM, Webhook, and other entry points can collaborate more stably
• Tool selection in multi-step tasks is more stable, and history replay is cleaner
4. IM and Webhook Experience Enhancements
When users initiate requests through IM entry points such as WeChat Work, Feishu (Lark), or DingTalk, they can receive instant “Received, processing” feedback more quickly. The Webhook side also supports configurable Shared Secrets, further enhancing system integration security.
• Ideal for real-time scenarios such as alerting, approvals, automated tickets, and enterprise bots
5. Management and Deployment Experience Optimization
Front-end management experience continues to be refined, covering session history, Runtime panels, role permission display, delete confirmations, and other details. On the deployment side, standardized runtime log output and log rotation configurations have been added, along with supplementary production runtime files for Docker, systemd, and logrotate.
• Facilitates long-term enterprise deployment, troubleshooting, maintenance, and auditing
Provider Updates: Truly Connecting Agents to Business Systems
Providers are the extension layer through which AtlasClaw connects to external systems and business capabilities. This round of Provider updates covers knowledge bases, cloud management, approvals, resource operations, R&D collaboration, and document artifact generation, enabling Agents to handle more real-world business tasks.
1. New Markdown Vault Local Knowledge Base Provider
A new markdown-vault Provider has been added, which can directly connect to local Markdown document libraries, supporting document parsing, knowledge retrieval, and Q&A based on local materials.
• Team handbooks, operations knowledge bases, project documentation, and internal Wikis can all become Agent knowledge sources
• No complex system modifications needed — existing document assets can be transformed into conversational, searchable knowledge capabilities
2. Enhanced SmartCMP Resource Request Capabilities
The SmartCMP Provider further optimizes the resource request pipeline, supporting multi-instance requests of the same type, multi-instance Schema-driven requests, and more reliable matching between services, resource packages, and parameter specifications.
• Users can describe resource needs in natural language, and the Agent assists in breaking down and generating more standardized cloud management platform requests
• Ideal for enterprise cloud management scenarios such as cloud resource requests, service catalog selection, and resource package matching
3. Enhanced SmartCMP Approval Capabilities
New pre-approval capabilities have been added, with optimized approval detail queries, to-do list display, specification name standardization, and approval output structure.
• Approvers can more quickly understand request content, view resource specifications, and process to-do items
• Request and approval workflows are no longer just form submissions — they can be understood and advanced with Agent assistance
4. Enhanced SmartCMP Resource Operations Capabilities
The resource operations side supports more follow-up operation scenarios, including parameter-free resource changes, resource evidence viewing, and selecting compliance analysis objects by name or list index.
• Alert cost optimization, resource compliance, and other skills can share a more complete resource view
• From discovering issues and analyzing risks to viewing evidence and triggering operations, a more complete operations closed loop is formed
5. Enhanced SmartCMP Authentication and Enterprise Integration
The SmartCMP Provider has enhanced Provider Token Auth and now supports Webhook Robot Execution Profiles.
• When external systems call AtlasClaw via Webhook, they can use clearer robot identities and credential boundaries to execute skills
• Ideal for automated workflows between alerting platforms, approval systems, and cloud management platforms
6. New GitHub Provider and General Document Capabilities
The Provider repository adds a GitHub Provider supporting Token-based authentication. It also adds text generation Skills, xlsx spreadsheet Skills, and optimizes workspace artifact exposure after presentation generation.
• R&D collaboration, Issue/PR management, and code repository operations now have the foundation for integration
• Text, tables, reports, presentations, and other work artifacts can be progressively incorporated into Agent workflows
Community and Resources
GitHub Main Repository: https://github.com/CloudChef/atlasclaw
Provider Plugin Repository: https://github.com/CloudChef/atlasclaw-providers
Project Website: https://atlasclaw.ai
Contact Email: atlasclaw@cloudchef.io
Star, Fork, open an Issue, or contribute a Provider plugin — join us in building a better enterprise AI Agent platform.
About Cloudchef
Cloudchef is a leading provider of cloud management and digital operations solutions, dedicated to helping enterprises accelerate their digital transformation through innovative cloud management platforms and services. With core products including the SmartCMP cloud management platform and AtlasClaw AI Agent, Cloudchef serves over 2,500 enterprise customers across finance, government, education, manufacturing, healthcare, and other industries. The company has been recognized as a Shanghai High-Tech Enterprise and has received multiple industry awards for its contributions to cloud management and IT operations innovation.