More on the program, including API credits and other benefits for venture-backed teams:

https://claude.com/programs/startups

Learn more about Salesforce in Claude: https://claude.com/blog/salesforce-in-claude

Evals call the model, so they use tokens and results vary. Pilot with `--runs 1` before a full run. Your plugin's hooks and MCP servers run as you, so only evaluate plugins you trust.

Run claude update to try it. Docs: https://code.claude.com/docs/en/plugin-evals

Then run `claude plugin eval`

You'll see each case's score with and without your plugin in your terminal, plus an HTML report with the full detail. If your account supports it, the report is also published as a private artifact.

Start in your plugin's folder and run `claude plugin eval init`

You tell Claude what good and bad output looks like and bring a few real prompts. Claude drafts the test cases and checks, pilots the suite, and tells you what a full run will cost.

We've also added auto mode to Claude Managed Agents.

With `auto`, Claude reviews each tool call based on your intent in `user.message` events and decides whether to run the tool call, deny it, or ask you for input.

We’re publishing this report so others can spot the same activity on their own platforms, and so we can give the public a clearer view of how emerging threats develop.

Read the report: https://www.anthropic.com/threat-intelligence-report-september-2026

These cases are not typical: we’re highlighting some of the most sophisticated misuse we’ve seen. But they’re especially important to discuss, because they show us where AI misuse is headed, where our safeguards work, and where they need to improve.

We disrupted every operation in the report, and used the lessons from them to strengthen our safeguards. Where appropriate, we also shared what we found with authorities and other AI companies.

We previously described some of the changes we’ve made to our alignment and security efforts following these incidents here: https://x.com/AnthropicAI/status/2094557124038951170

Our initial agreement runs for eight weeks, and we intend to give METR as much time as it deems necessary to complete a thorough investigation. https://www.anthropic.com/research/alignment-assessment-cybersecurity-incidents

METR will also conduct an independent investigation, with wide-ranging access, including to transcripts beyond the window in which the incidents occurred, and to Anthropic employees permitted to share confidential information.

If you're building on the Claude Platform, get listed and make it easier for our customers to buy your product: https://claude.com/marketplace-partners

Like all economic models, ours simplifies a more complex reality. But by building better scenarios of our possible economic future, we can take steps to make sure everyone benefits from it.