The QIR documentation pairs a wide robotics catalog with the less visible rules that package components, gather dependencies and set a ROS 2 build configuration.

2 min

Sep 4, 2026Tools

Research Tools

Databricks Says Proteus Generated Qwen Kernels Up to 5.2× Faster Than vLLM

Proteus specializes GPU code for particular runtime shapes rather than seeking one universal kernel. Databricks’ results highlight both the potential of that approach and the validation work needed to trust it.

3 min

Sep 4, 2026Tools

Launch Tools

GitHub Launches HydraFusion, Citing 67% Lower Cost on Terminal-Bench 2.1

The Copilot research preview turns model selection into a per-task workflow. Its strongest published savings result is a GitHub-run, tuned offline evaluation—not a measure of ordinary repository work.

2 min

Sep 4, 2026Tools

Launch Tools

xAI Publishes Procurement Bot Prompt That Requires Approval for Vendor Messages—and Bars Purchases

The Haggle Bot template turns vendor savings into a persistent agent workflow: inspect live spend and usage, price alternatives, and prepare negotiations while keeping external outreach and financial commitments behind separate limits.

3 min

Sep 3, 2026Tools

Tools

Anthropic Has Packaged Android App Testing for Claude Code but Kept It Disabled

The packaged workflow would let Claude run Android apps, move through screens, and capture evidence in emulators. Anthropic has not said when it will turn the capability on.

2 min

Sep 3, 2026Tools

Launch Tools

claude-api-guard Scans AI SDK Changes in CI, but Misses Raw HTTP Calls

The GitHub Action can block known, high-severity SDK compatibility problems before a merge and fix some mechanical changes. Its coverage remains tied to recognizable official SDK patterns and a limited set of languages.

2 min

Sep 3, 2026Tools

Acquisition Tools

Plan A Acquires AI Tool to Speed Technical Due Diligence for M&A

The acquired system is intended to connect signals across complex technology estates before a deal closes. Its practical value will depend on integration and performance details Plan A has not disclosed.

3 min

Sep 3, 2026Tools

Launch Tools

Anthropic Adds ant apply to Manage Claude API Resources as Code

The new CLI command lets teams keep Claude API resource definitions alongside application code, while its lockfile sets strict limits on what the tool can manage.

2 min

Sep 3, 2026Tools

Launch Tools

Proofpoint Adds an OpenAI-Powered Security Agent That Cannot Contain Threats

The SOC Analyst Agent assembles evidence and recommends next steps across Proofpoint’s products, but consequential response actions remain with security staff.

2 min

Sep 3, 2026Tools

Tools

Oracle Publishes Agent Design That Keeps Credentials and Proof Outside the Model

Oracle’s proposed harness treats a model’s output as a request for action, leaving systems outside the model to authorize work, retain state and verify its result.

2 min

Sep 3, 2026Tools

Research Tools

Cohere Publishes Agent-Tool Dataset, Finds 419 Occupations Have No Coverage

The new ATE dataset offers a supply-side map of what developers have made available for AI agents, not proof of workplace adoption. Its central finding is less about how many tools exist than which job tasks they reach—and which they leave alone.

3 min

Sep 3, 2026Tools

Partnership Tools

Perplexity Adds One-Command API Setup Through Stripe Projects

The terminal flow puts credentials, prepaid credits and spending controls beside an application’s code. But teams connecting an existing Perplexity project stay on its existing billing path.

2 min

Sep 2, 2026Tools

Open release Tools

Qdrant Publishes a 10B-Vector Benchmark and Open Tools to Rerun It

FineWeb-10B pairs an unusually large retrieval corpus with a public workflow for building and testing it. The useful test now is whether other database operators can reproduce comparisons on the same workload.

2 min

Sep 2, 2026Tools

Partnership Tools

LangChain Adds Agent Payments With Spending Caps, but Settlement Questions Remain

The Nevermined integration lets an agent purchase data and API access during a task under delegated-card limits. An open repository issue, however, raises a separate question about whether some paid results can be released when settlement fails.

3 min

Sep 2, 2026Tools

Security risk Tools

Capsule Security Adds a Last-Second Block for AI-Agent Actions

The new control layer judges a planned action before it runs, rather than relying only on access permissions or post-incident review. Its reported benchmark results are promising, but the company’s internal model comparison remains difficult to assess fully.

2 min

Sep 2, 2026Tools

Benchmark Tools

LlamaIndex and Kaggle Launch a Document Benchmark for Cost and Long Files

ExtractBench publishes a broader test for enterprise document extraction, while its launch results include company-reported figures for LlamaIndex’s own product.

2 min

Sep 2, 2026Tools

Open release Tools

Oracle Publishes AI Agent Blueprint That Keeps Database Writes Under Human Approval

Oracle’s supply-chain sample connects chat interfaces to inventory recommendations, but limits execution to a procedure-backed, audited database path after a planner approves it.

2 min

Sep 2, 2026Tools

Launch Tools

Google Launches Fairwind to Give Selected Defenders AI Tools for Finding and Fixing Flaws

The program draws a line between broadly available code-security tooling and restricted access to Gemini 3.8 Flash Cyber, reflecting the dual-use risk of AI that can identify and repair software weaknesses.

3 min

Load more stories18 more next