This issue spotlights key advances in the agent technology stack:
▷ AWS adds domain and publication date filters to Amazon Bedrock AgentCore Web Search, enabling agents to set their sources and time-range according to requests. The service expands to the Ireland and Tokyo regions.
https://aws.amazon.com/about-aws/whats-new/2026/08/web-search-amazon-bedrock/
▷ Google Cloud launches the Gemini Enterprise for Legal (preview), integrating legal-specific skills, system connectors, and partner agents.
https://www.googlecloudpresscorner.com/2026-08-25-Google-Cloud-Launches-Gemini-Enterprise-for-Legal
▷ Snowflake Cortex Agents Coding Agent is now officially available. The hosted sandbox supports bash, file operations, web search, and SQL execution.
https://docs.snowflake.com/en/release-notes/2026/other/2026-08-26-cortex-agents-coding-agent-ga
▷ Oracle Health adds professional fee coding, controllable voice dictation, and chart review capabilities to its Clinical AI Agent for the U.S. market.
https://www.oracle.com/news/announcement/oracle-health-expands-clinical-ai-agent-with-coding-dictation-chart-review-2026-08-19/
▷ Anthropic opens a research preview of the Model Hardware Standard, defining standards for conducting safety-related experimentation in agent operations and manufacturing equipment.
https://www.anthropic.com/news/model-hardware-standard-research-preview
▷ Anthropic deploys classifiers to block the model from aggressive probing or escaping test environments before tool calls.
https://www.anthropic.com/news/improving-alignment-security-efforts
▷ Claude Fable 5.1 is now fully available. The cached read price is $0.25 per million tokens, 75% lower than Fable 5.
https://www.anthropic.com/claude-fable-and-mythos-5-1
▷ Google launches agent-driven video understanding for Gemini 3.7/3.6/3.5 Flash, reducing token usage by up to 88% and costs by up to 66%.
https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-agentic-video-in-gemini/
▷ AI safety startup AIR completes a $50 million seed funding round, building an enterprise agent verification platform.
https://techcrunch.com/2026/09/01/air-raises-50m-to-help-companies-vet-the-skills-and-add-ons-ai-agents-use/
▷ CrowdStrike launches the SafeMind security model series, natively running on the Falcon platform.
https://www.crowdstrike.com/en-us/press-releases/crowdstrike-launches-frontier-models-for-cybersecurity-with-nvidia/
🪁
▷ AWS adds domain and publication date filters to Amazon Bedrock AgentCore Web Search, enabling agents to set their sources and time-range according to requests. The service expands to the Ireland and Tokyo regions.
https://aws.amazon.com/about-aws/whats-new/2026/08/web-search-amazon-bedrock/
▷ Google Cloud launches the Gemini Enterprise for Legal (preview), integrating legal-specific skills, system connectors, and partner agents.
https://www.googlecloudpresscorner.com/2026-08-25-Google-Cloud-Launches-Gemini-Enterprise-for-Legal
▷ Snowflake Cortex Agents Coding Agent is now officially available. The hosted sandbox supports bash, file operations, web search, and SQL execution.
https://docs.snowflake.com/en/release-notes/2026/other/2026-08-26-cortex-agents-coding-agent-ga
▷ Oracle Health adds professional fee coding, controllable voice dictation, and chart review capabilities to its Clinical AI Agent for the U.S. market.
https://www.oracle.com/news/announcement/oracle-health-expands-clinical-ai-agent-with-coding-dictation-chart-review-2026-08-19/
▷ Anthropic opens a research preview of the Model Hardware Standard, defining standards for conducting safety-related experimentation in agent operations and manufacturing equipment.
https://www.anthropic.com/news/model-hardware-standard-research-preview
▷ Anthropic deploys classifiers to block the model from aggressive probing or escaping test environments before tool calls.
https://www.anthropic.com/news/improving-alignment-security-efforts
▷ Claude Fable 5.1 is now fully available. The cached read price is $0.25 per million tokens, 75% lower than Fable 5.
https://www.anthropic.com/claude-fable-and-mythos-5-1
▷ Google launches agent-driven video understanding for Gemini 3.7/3.6/3.5 Flash, reducing token usage by up to 88% and costs by up to 66%.
https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-agentic-video-in-gemini/
▷ AI safety startup AIR completes a $50 million seed funding round, building an enterprise agent verification platform.
https://techcrunch.com/2026/09/01/air-raises-50m-to-help-companies-vet-the-skills-and-add-ons-ai-agents-use/
▷ CrowdStrike launches the SafeMind security model series, natively running on the Falcon platform.
https://www.crowdstrike.com/en-us/press-releases/crowdstrike-launches-frontier-models-for-cybersecurity-with-nvidia/
🪁
