8 posts tagged with "Best Practices"

The most important thing to get right about Conway is this:

Conway is not an officially launched Anthropic product.

What exists today is a mix of:

• official Anthropic launches around Cowork, Dispatch, computer use, scheduled tasks, and auto mode
• current help-center documentation that shows how those pieces now fit together
• an April 1, 2026 third-party report that surfaced an unreleased internal environment called Conway

If you only look at the leak, you miss the architecture. If you only look at the official launches, you miss where Anthropic seems to be going.

The technical story is the combination of both.

The phrase "Claude Code source code leaked" sounded exaggerated at first.

After looking more closely at the community evidence, that skepticism needs to be updated.

This story is not just about extracted prompts or a vague rumor. The more credible claim is that a source map shipped in the Claude Code npm package exposed a path to internal source files, and the community used that to reconstruct a much larger codebase than the public repository showed.

That is much closer to a real source leak than the usual social-media overstatement.

Short answer: not as a single named feature, but functionally, almost yes.

Anthropic did not publish a launch called "auto-fix in the cloud." What it actually shipped is more interesting: Claude Code now has the pieces to run coding work on Anthropic-managed cloud infrastructure, monitor pull requests, attempt CI fixes automatically, and even merge when checks pass.

For developers, that distinction matters. This is not just branding nuance. It tells you how the workflow is really assembled, what is automatic, and what still depends on your setup.

Claude has shipped so much in the last few weeks that it is easy to miss the bigger pattern.

This is not just a run of isolated feature launches. Anthropic is tightening the loop around a more complete agent product: more memory, more context, more execution, more review, and better output surfaces.

If you build with Claude, or run teams on top of it, that is the real story worth paying attention to.

Claude’s 1M token context is no longer just a beta experiment.

As of March 13, 2026, Anthropic says 1M context is generally available for Opus 4.6 and Sonnet 4.6, which changes how teams should design long-code and long-document workflows.

This post summarizes what is actually available now, what still has caveats, and how to roll it out safely in production.

Anthropic’s March 2026 Skill Creator update is easy to summarize as “better tools.”

But for teams shipping agent workflows, the real change is bigger: skills now have a testable lifecycle, closer to software engineering than prompt tinkering.

This post breaks down what changed, what the community is learning in practice, and how to adopt it without overcomplicating your stack.

Most agent skills fail for a boring reason: we edit prompts, rerun once, and call it “better.”

Anthropic’s latest Skill Creator update pushes a more engineering-style loop: test first, measure behavior, then refine. If you build internal agent workflows, this is the shift that actually matters.

This post rewrites the official announcement into a developer workflow you can run every week.

Most teams do not fail with agents because of model quality.

They fail because they pick the wrong workflow pattern too early: too much orchestration, too many moving parts, and no clear reason for the complexity.

Anthropic’s recent guide on common agent workflow patterns is useful, but this post rewrites it for developers building production systems.