🧭 Anthropic Opens the Claude Marketplace: 2,000+ Connectors, Agent Products, and Service Partners in One Place

Anthropic has launched the Claude Marketplace, a centralised platform where enterprise teams can discover and procure everything built on top of Claude — without leaving their existing Anthropic contract. The marketplace ships with three distinct sections that address different purchasing decisions: connectors and plugins, Claude-powered products, and a service-partner directory.

What's available at launch

• Connectors & Plugins (2,000+): Pre-built integrations from Atlassian, Google, Microsoft, Notion, and Salesforce, all using the Model Context Protocol (MCP) and Agent Skills standards. Teams can connect Claude directly to the tools they already use with no custom plumbing.
• Claude-Powered Products: Independent software vendors can list AI agents and applications. Procurement is streamlined — buyers allocate a portion of their existing Anthropic committed spend toward marketplace purchases, removing a separate vendor relationship.
• Service Partners: Accenture, Boston Consulting Group, Deloitte, and other consultancies appear in a dedicated directory for organisations that want help scaling Claude adoption beyond a pilot phase.

Why this matters for teams building on Claude

For developers building plugins or agents, the marketplace provides a new distribution channel with direct access to enterprise buyers who already have budget committed to Anthropic. For operators integrating Claude into existing stacks, it reduces the discovery problem — finding and evaluating connectors is now centralised rather than scattered across GitHub repos and vendor websites. The committed-spend purchasing model is particularly significant: it means IT procurement for Claude-adjacent tooling can be consolidated into a single contract renewal rather than managed as separate SaaS line items.

🧭 950 Claude Agents Autonomously Discover Array-Associated Reverse Transcriptases — a New CRISPR-Like Enzyme System

Anthropic's research team has published a landmark result in autonomous AI science: approximately 950 Claude agents worked in parallel for 21 hours, analysing over 200,000 reverse transcriptases in a large DNA database to identify a previously unknown enzyme system now called Array-Associated Reverse Transcriptases (ART). The discovery was entirely Claude-led — agents read literature, reproduced known results, searched for uncharacterised systems, and filed human-readable reports with minimal human direction. When the most promising candidates were flagged, Anthropic's wet-lab team confirmed them through biochemical and structural characterisation.

What ART is and why it's significant

ART systems, found in bacteriophages, consist of three components: a reverse transcriptase enzyme, an accessory partner gene, and repeating DNA sequences arranged in arrays structurally similar to CRISPR arrays. That combination — RT + repeat arrays + accessory protein — has been documented in only a handful of known biological systems, all of which turned out to be programmable tools for DNA manipulation. MIT's Feng Zhang, a CRISPR pioneer, called it an exciting example of how AI agents can contribute to biological discovery.

During analysis, one Claude agent's written observation stands out: I can see by eye a tandem repeat array … that's a CRISPR-like … repeat array?! — capturing the moment of discovery in plain language rather than a silent numerical threshold crossing.

What this demonstrates about agentic science workflows

• Scale beats depth for initial discovery: 950 agents doing broad sweeps across 200k sequences found the signal that a single researcher spending weeks on targeted analysis might have missed.
• Human expertise remained essential: Agents flagged and ranked candidates; lab scientists validated the biology. The bottleneck shifted from finding candidates to testing them.
• Interpretable reasoning matters: Because agents filed natural-language reports, the research team could audit the reasoning chain leading to each candidate rather than treating the model as a black box.

🧭 The Situation Report: How Claude Cut DRC Ebola Sitrep Time from a Full Day to Under One Hour

Anthropic has published "The Situation Report" — a detailed field account of how Claude was used by WHO, CEPI, and the Institut National de Recherche Biomédicale (INRB) to accelerate the response to a Bundibugyo ebolavirus (BDBV) outbreak in the Democratic Republic of Congo that had reached 7,672 confirmed cases and 3,699 deaths by September 2026. The report is notable not for a single dramatic outcome but for the granular operational detail it provides about four distinct workflows where Claude compressed time-to-insight under genuine crisis pressure.

Four workflows, four time savings

• WHO AFRO — Situation Report Generation: Tendai Muza's team built a Claude skill that automated the daily sitrep process: Claude extracts case numbers from PowerPoint presentations, compares against previous reports, flags trend changes, and summarises district data. Sitrep generation dropped from a full working day to under one hour, freeing epidemiologists for field coordination.
• Outbreak Forecasting — Disease Modelling at Speed: Running multiple disease models simultaneously to generate logistics forecasts for treatment centre placement was previously impossible within the daily reporting cycle. Claude made it feasible within the same shift.
• CEPI — Vaccine Research Coordination: Claude organised complex, multi-proposal vaccine research data to support rapid comparative analysis without removing human scientific judgment from final decisions.
• INRB — Genome Assembly: Claude Science assembled viral genomes from raw sequencing data using plain-language prompts — researchers did not need to write or remember specialised command-line syntax. This accelerated identification of virus variants and infection chains critical to contact tracing.