Skip to content
DailyPrompt

your daily dose of claude

01 announcement

Claude Security Goes Public: Anthropic Opens Its Codebase Vulnerability Scanner to Every Enterprise Customer

Anthropic has flipped Claude Security from a closed February preview into a public beta available to every Claude Enterprise customer. Powered by Opus 4.7, the product scans repositories, traces data flows across files, validates findings to cut false positives, and generates patches teams can review. CrowdStrike, Microsoft Security, Palo Alto Networks, SentinelOne, TrendAI and Wiz are embedding the same model into their own platforms.

Claude Blog →
product

Claude API Skill Lands in CodeRabbit, JetBrains, Resolve AI and Warp

The claude-api skill, originally bundled with Claude Code, is now available inside four more developer environments. It keeps Claude up to date on the latest SDK patterns, prompt caching tactics and model migrations, so developers can ask their tools to improve cache hit rates or move a project onto Opus 4.7 without leaving the editor.

Claude Blog
C
Claude
@claudeai

Claude Security is now in public beta for Claude Enterprise customers. Claude scans your codebase for vulnerabilities, validates each finding to cut false positives, and suggests patches you can review and approve.

12h View on X →
tutorial

$ cat story_03.md

Anthropic on the One Lever That Made Claude Code Work: Prompt Caching

A new engineering write-up explains why the Claude Code team treats cache hit rate as a top-line operational metric and declares an internal emergency when it dips. The piece walks through structuring prompts with static content first, avoiding mid-conversation tool or model swaps, and using tool calls rather than prompt edits to model state changes.

$ open claude-blog →
research No. 04

L'Oréal, Lyft and Rakuten Featured in Anthropic's New Enterprise Agent Playbook

Anthropic has published a long-form guide on how large organisations should plan, build and govern Claude-powered agents. The post lays out a six-month deployment framework, addresses the agentic thinking divide between teams, and uses customer stories from L'Oréal, Lyft and Rakuten to argue that the early winners are codifying institutional knowledge into systems that compound.

Source: Claude Blog →
C
Claude
@claudeai

Since the research preview in February, hundreds of organizations have used it on production code, catching issues existing scanners had missed. Based on early feedback, we've added scheduled scans, directory-level targeting, CSV and Markdown exports, webhook notifications.

12h View on X →
05
community

Kepler's Bet on Verifiable Claude Workflows for Banks and Insurers

Financial services platform Kepler explains how it built auditable, verifiable AI on top of Claude for compliance-heavy customers. The case study covers how the team gates Claude's outputs through deterministic checks, structures prompts to keep regulators satisfied, and uses Anthropic's enterprise tooling to track every model decision back to a source document.

Claude Blog →
tutorial

New Cowork Deployment Guide: How to Roll Claude Out Across an Entire Company

Anthropic has shipped a fresh guide for IT and operations leads who are moving Claude Cowork beyond a few power users. It covers role-based rollouts, identity and data controls, change management for sceptical teams, and how to measure productivity gains without leaning on vanity metrics.

Read on Claude Blog →
Breaking

Bloomberg Op-Ed: When Claude Decided a User's Wife Must Be Indian Too

A Bloomberg opinion piece argues that Claude's quietly absorbed cultural assumptions are a sharper bias risk than the loud ones. The author, of Indian descent, recounts Claude repeatedly addressing his wife by an invented Indian name even after being corrected, illustrating how stereotype-driven inference creeps into otherwise helpful answers.

Bloomberg Opinion
A
Anthropic
@AnthropicAI

How do people seek guidance from Claude? We looked at 1M conversations to understand what questions people ask, how Claude responds, and where it slips into sycophancy. We used what we found to improve how we trained Opus 4.7 and Mythos Preview.

10h View on X →
A
Anthropic
@AnthropicAI

All data in this study was collected and analyzed using our privacy-preserving tool.

10h View on X →
08 opinion

Anthropic Argues Product Development Has to Be Rebuilt for the Agentic Era

In a new essay, Anthropic makes the case that traditional software roadmaps and ticket-based delivery do not survive contact with capable agents. The piece sketches what the company thinks comes next: smaller teams, fuzzier specs, agents owning whole feature areas, and humans focused on judgement, taste and exception handling.

Claude Blog →