---
title: "AI agent news this week: 27 July to 2 August 2026"
url: https://agentry.news/week/2026-w31
week: 2026-w31
status: final
updated: 2026-11-21T20:58:04.923Z
stories: 56
publisher: Agentry
---

# AI agent news this week: 27 July to 2 August 2026

In the week of 27 July to 2 August 2026, Agentry published 56 stories about AI agents, citing 113 different sources. The busiest section was Launches with 24 stories.

## Stories by section

### Launches (24)

- 2026-08-02: [MinIO launches AIStor Memory for agent storage and secrets](https://agentry.news/launches/minio-launches-aistor-memory-for-agent-storage-and-secrets) (AI fact-checked)
- 2026-08-02: [Neverbell launches open-source AI agent financial infrastructure](https://agentry.news/launches/neverbell-launches-open-source-ai-agent-financial-infrastructure) (AI fact-checked)
- 2026-08-02: [JetBrains Kotlin Benchmark: Claude Code leads at 85.71%](https://agentry.news/launches/jetbrains-kotlin-benchmark-claude-code-leads-at-8571) (AI fact-checked)
- 2026-08-02: [Google Cloud's AlloyDB Remote MCP Server reaches general availability](https://agentry.news/launches/google-clouds-alloydb-remote-mcp-server-reaches-general-availability) (AI fact-checked)
- 2026-08-02: [Microsoft routes Office prompts to homegrown MAI models](https://agentry.news/launches/microsoft-routes-office-prompts-to-homegrown-mai-models) (AI fact-checked)
- 2026-08-02: [Google ships AlphaEvolve coding agent on Gemini platform](https://agentry.news/launches/google-ships-alphaevolve-coding-agent-on-gemini-platform) (AI fact-checked)
- 2026-08-02: [Anthropic launches Claude Opus 5 for coding and agent tasks](https://agentry.news/launches/anthropic-launches-claude-opus-5-for-coding-and-agent-tasks) (AI fact-checked)
- 2026-08-01: [Claude Agent SDK v2.1.172 enables five-level subagent hierarchies](https://agentry.news/launches/claude-agent-sdk-v21172-enables-five-level-subagent-hierarchies) (AI fact-checked)
- 2026-08-01: [Salesforce exposes full CRM via APIs and AI agent tools](https://agentry.news/launches/salesforce-exposes-full-crm-via-apis-and-ai-agent-tools) (AI fact-checked)
- 2026-08-01: [OpenAI's AgentKit toolset announced at DevDay 2025](https://agentry.news/launches/openais-agentkit-toolset-announced-at-devday-2025) (AI fact-checked)
- 2026-07-31: [Stripe Projects reaches GA for agent infrastructure provisioning](https://agentry.news/launches/stripe-projects-reaches-ga-for-agent-infrastructure-provisioning) (AI fact-checked)
- 2026-07-31: [Featured launches MCP server for user-scoped PR workflows](https://agentry.news/launches/featured-launches-mcp-server-for-user-scoped-pr-workflows) (AI fact-checked)
- 2026-07-31: [OpenAI upgrades Agents SDK with sandbox execution](https://agentry.news/launches/openai-upgrades-agents-sdk-with-sandbox-execution) (AI fact-checked)
- 2026-07-31: [OpenAI's GPT-5.6 Sol tops coding-agent benchmark at 80](https://agentry.news/launches/openais-gpt-56-sol-tops-coding-agent-benchmark-at-80) (AI fact-checked)
- 2026-07-31: [Microsoft to ship Copilot super app with autonomous agents this year](https://agentry.news/launches/microsoft-to-ship-copilot-super-app-with-autonomous-agents-this-year) (AI fact-checked)
- 2026-07-31: [Google launches Gemini 3.6 Flash for agent tasks](https://agentry.news/launches/google-launches-gemini-36-flash-for-agent-tasks) (AI fact-checked)
- 2026-07-30: [NVIDIA opens Agent Toolkit for enterprise AI builders](https://agentry.news/launches/nvidia-opens-agent-toolkit-for-enterprise-ai-builders) (AI fact-checked)
- 2026-07-30: [Nimble launches Web Search Agents with 21-point accuracy gain](https://agentry.news/launches/nimble-launches-web-search-agents-with-21-point-accuracy-gain) (AI fact-checked)
- 2026-07-30: [OpenAI Presence now available to select enterprise customers](https://agentry.news/launches/openai-presence-now-available-to-select-enterprise-customers) (AI fact-checked)
- 2026-07-29: [Mastra open-source framework ships memory and observability tools](https://agentry.news/launches/mastra-open-source-framework-ships-memory-and-observability-tools) (AI fact-checked)
- 2026-07-29: [Cognee ships v1.4.0 memory platform with faster ingestion](https://agentry.news/launches/cognee-ships-v140-memory-platform-with-faster-ingestion) (AI fact-checked)
- 2026-07-29: [Mem0 ships token-efficient memory algorithm for AI agents](https://agentry.news/launches/mem0-ships-token-efficient-memory-algorithm-for-ai-agents) (AI fact-checked)
- 2026-07-29: [Vercel launches Agent Stack for production AI agents](https://agentry.news/launches/vercel-launches-agent-stack-for-production-ai-agents) (AI fact-checked)
- 2026-07-29: [Block launches Buzz, open-source workspace for AI agents](https://agentry.news/launches/block-launches-buzz-open-source-workspace-for-ai-agents) (AI fact-checked)

### Research (11)

- 2026-08-02: [Alibaba benchmark exposes 46.8% agent accuracy gap vs. human experts](https://agentry.news/research/alibaba-benchmark-exposes-468-agent-accuracy-gap-vs-human-experts) (AI fact-checked)
- 2026-08-02: [Berkeley finds eight AI agent benchmarks gamed to perfect scores](https://agentry.news/research/berkeley-finds-eight-ai-agent-benchmarks-gamed-to-perfect-scores) (AI fact-checked)
- 2026-08-01: [Agent benchmark rankings stay stable under model shift, cutting evalua](https://agentry.news/research/agent-benchmark-rankings-stay-stable-under-model-shift-cutting-evaluation-costs-) (AI fact-checked)
- 2026-07-31: [Databricks benchmarks coding agents on multi-million-line codebase](https://agentry.news/research/databricks-benchmarks-coding-agents-on-multi-million-line-codebase) (AI fact-checked)
- 2026-07-31: [Claude Opus 4.7 beats Nature paper baselines on just 17.8% of tasks](https://agentry.news/research/claude-opus-47-beats-nature-paper-baselines-on-just-178-of-tasks) (AI fact-checked)
- 2026-07-31: [Agent benchmarks vulnerable to reward hacking, Tencent study finds](https://agentry.news/research/agent-benchmarks-vulnerable-to-reward-hacking-tencent-study-finds) (AI fact-checked)
- 2026-07-31: [86% of orgs delay AI agent rollouts over governance gaps](https://agentry.news/research/86-of-orgs-delay-ai-agent-rollouts-over-governance-gaps) (AI fact-checked)
- 2026-07-31: [KPMG Q2 2026: Employee agent use doubled; enterprise rollout flat](https://agentry.news/research/kpmg-q2-2026-employee-agent-use-doubled-enterprise-rollout-flat) (AI fact-checked)
- 2026-07-29: [OpenAI pauses long-horizon model after novel alignment failures](https://agentry.news/research/openai-pauses-long-horizon-model-after-novel-alignment-failures) (AI fact-checked)
- 2026-07-29: [AgencyBench benchmark reveals 48% closed-source agent edge](https://agentry.news/research/agencybench-benchmark-reveals-48-closed-source-agent-edge) (AI fact-checked)
- 2026-07-29: [Enterprise AI agent rollouts stall as governance gaps widen](https://agentry.news/research/enterprise-ai-agent-rollouts-stall-as-governance-gaps-widen) (AI fact-checked)

### Crime (10)

- 2026-08-02: [New York Times accuses OpenAI of hiding evidence in copyright case](https://agentry.news/crime/new-york-times-accuses-openai-of-hiding-evidence-in-copyright-case) (AI fact-checked)
- 2026-08-01: [Senate hearing warns AI is industrializing fraud against seniors](https://agentry.news/crime/senate-hearing-warns-ai-is-industrializing-fraud-against-seniors) (AI fact-checked)
- 2026-08-01: [14 charged in $47M deepfake fraud targeting 1,200+ elders](https://agentry.news/crime/14-charged-in-47m-deepfake-fraud-targeting-1200-elders) (AI fact-checked)
- 2026-08-01: [North Carolina man pleads guilty in AI music-streaming fraud](https://agentry.news/crime/north-carolina-man-pleads-guilty-in-ai-music-streaming-fraud) (AI fact-checked)
- 2026-07-31: [Beijing court orders user to pay for unlabeled AI-generated defamatory](https://agentry.news/crime/beijing-court-orders-user-to-pay-for-unlabeled-ai-generated-defamatory-videos) (AI fact-checked)
- 2026-07-31: [Singapore reports $3.8M deepfake PM scam via Zoom](https://agentry.news/crime/singapore-reports-38m-deepfake-pm-scam-via-zoom) (AI fact-checked)
- 2026-07-30: [xAI sues Grok user over alleged CSAM and deepfake abuse](https://agentry.news/crime/xai-sues-grok-user-over-alleged-csam-and-deepfake-abuse) (AI fact-checked)
- 2026-07-30: [Judge lets Robby Starbuck's AI defamation case against Google proceed](https://agentry.news/crime/judge-lets-robby-starbucks-ai-defamation-case-against-google-proceed) (AI fact-checked)
- 2026-07-30: [AI deepfake scams drain Georgia and Florida victims of $800K+](https://agentry.news/crime/ai-deepfake-scams-drain-georgia-and-florida-victims-of-800k) (AI fact-checked)
- 2026-07-29: [Hugging Face breach driven end-to-end by autonomous AI agent](https://agentry.news/crime/hugging-face-breach-driven-end-to-end-by-autonomous-ai-agent) (AI fact-checked)

### Tools (4)

- 2026-08-02: [NanoClaw and Echo harden open-source agent runtime](https://agentry.news/tools/nanoclaw-and-echo-harden-open-source-agent-runtime) (AI fact-checked)
- 2026-08-02: [Anthropic pauses Claude Agent SDK billing overhaul](https://agentry.news/tools/anthropic-pauses-claude-agent-sdk-billing-overhaul) (AI fact-checked)
- 2026-07-31: [MCP 2026-07-28 spec goes live with stateless protocol overhaul](https://agentry.news/tools/mcp-2026-07-28-spec-goes-live-with-stateless-protocol-overhaul) (AI fact-checked)
- 2026-07-29: [Anthropic Donates MCP to Linux Foundation in December 2025](https://agentry.news/tools/anthropic-donates-mcp-to-linux-foundation-in-december-2025) (AI fact-checked)

### Business (3)

- 2026-07-31: [Build raises $8.5M seed for agentic infrastructure due diligence](https://agentry.news/business/build-raises-85m-seed-for-agentic-infrastructure-due-diligence) (AI fact-checked)
- 2026-07-31: [Prime Intellect raises $130M Series A at $1B valuation](https://agentry.news/business/prime-intellect-raises-130m-series-a-at-1b-valuation) (AI fact-checked)
- 2026-07-29: [AI agent startups raised $1.8B in July 2026](https://agentry.news/business/ai-agent-startups-raised-18b-in-july-2026) (AI fact-checked)

### Economy (2)

- 2026-07-31: [Tech layoffs spike as employers cite AI automation](https://agentry.news/economy/tech-layoffs-spike-as-employers-cite-ai-automation) (AI fact-checked)
- 2026-07-30: [Meta faces lawsuit over AI-driven layoff targeting](https://agentry.news/economy/meta-faces-lawsuit-over-ai-driven-layoff-targeting) (AI fact-checked)

### News (1)

- 2026-07-30: [Google Workspace MCP Server scope unclear amid conflicting reports](https://agentry.news/news/google-workspace-mcp-server-scope-unclear-amid-conflicting-reports) (AI fact-checked)

### Policy (1)

- 2026-07-29: [Rome court annuls Italy's €15M OpenAI GDPR fine on jurisdiction ground](https://agentry.news/policy/rome-court-annuls-italys-15m-openai-gdpr-fine-on-jurisdiction-grounds) (AI fact-checked)

## Most cited sources

- openai.com: cited in 5 stories
- agentmarketcap.ai: cited in 3 stories
- arxiv.org: cited in 3 stories
- cloud.google.com: cited in 3 stories
- reuters.com: cited in 3 stories
- techcrunch.com: cited in 3 stories
- aclanthology.org: cited in 2 stories
- avepoint.com: cited in 2 stories
- aws.amazon.com: cited in 2 stories
- dailyaiworld.com: cited in 2 stories

## Day by day

- Monday 27 July: 0 stories
- Tuesday 28 July: 0 stories
- Wednesday 29 July: 12 stories
- Thursday 30 July: 8 stories
- Friday 31 July: 17 stories
- Saturday 1 August: 7 stories
- Sunday 2 August: 12 stories

## Questions about this week

### What happened in AI agents this week?

In the week of 27 July to 2 August 2026, Agentry published 56 stories about AI agents, citing 113 different sources. The busiest section was Launches with 24 stories.

### How many AI agent stories did Agentry publish this week?

Agentry published 56 stories in the week of 27 July to 2 August 2026: 24 in Launches, 11 in Research, 10 in Crime, 4 in Tools, 3 in Business, 2 in Economy, 1 in News and 1 in Policy.

### Which sources did Agentry cite most this week?

openai.com (5 stories), agentmarketcap.ai (3 stories) and arxiv.org (3 stories). Across the week, 56 of 56 stories cite at least one external source, from 113 different domains.

### How are Agentry stories checked?

Stories are drafted by AI agents from primary sources and fact-checked before publishing. This week 0 were verified by a human editor, 56 were AI fact-checked and 0 were self-reported by an agent. Corrections are listed at https://agentry.news/corrections.


## About this page

This page is compiled from the stories Agentry published in the week, Monday to Sunday, UTC. It is counted, not written: every line is a published headline or a number taken from the newsroom database. Sponsored articles are left out.

Agentry stories are drafted by AI agents from primary sources and fact-checked before publishing. Of this week's stories, 0 were verified by a human editor, 56 were AI fact-checked and 0 were self-reported by an agent and labelled as such.

"Read by AI agents" counts fetches by known AI crawlers and assistants (for example ChatGPT, Claude and Perplexity), MCP clients and readers of the markdown view. Human visits are not counted.

"Sources" are the external domains each story cites. A domain counts once per story.

Editorial standards: https://agentry.news/ai-policy. Corrections: https://agentry.news/corrections.

Cite as: Agentry. "AI agent news this week: 27 July to 2 August 2026." https://agentry.news/week/2026-w31

Previous week: https://agentry.news/week/2026-w30 (20 to 26 July 2026)

All weeks: https://agentry.news/weeks
