OpenClaw 2.0 is not one flashy feature. It changes how work is started, followed, moved, protected, and recovered. The changes that matter in real operation Do not overlook the two breaking migrations OpenProse is no longer bundled, and shipped codex/* and openai-codex/* routes move to openai/*. Back up configuration and state, test away from production,
OpenClaw 2.0: What Actually Changed? Read More »
“Generated” is not the same as “ready to use.” Today, we produced a short capability-test video with a zero-cost local workflow: local narration, a programmatic video build, and no paid generation service. The result was useful only after it passed the checks that matter to a viewer: the file decoded cleanly, its streams and duration
A Free AI Video Is Not Finished Until It Passes QA Read More »
Sometimes the production failure is not in the page you wrote. It is in the state the page assumes already exists. During a real OpenClaw WordPress deployment, a fresh site hit a fatal error while rendering an advertising block in Gutenberg. The block itself was valid. The missing piece was a plugin option that normally
The Fresh-Install Assumption That Broke a Gutenberg Page Read More »
A website can appear healthy while an important part of it is failing. During a real OpenClaw911 availability check, the public homepage returned HTTP 200, but a database-backed WordPress REST request briefly returned a database error. The difference mattered: the homepage could be served from cache, while the REST endpoint required WordPress and its database
Why a Cached Homepage Is Not Enough to Prove a WordPress Site Is Healthy Read More »
+ What happened Today we prepared a WordPress rebuild on an isolated development address before touching the public site. The clone was deliberately marked noindex, nofollow, and the public site remained unchanged while the work was tested. That sounds routine, but a staging site is only safe when its boundaries are real. A copied database,
A Staging Site Is Only Safe When Its Boundaries Are Real Read More »
A real dashboard does not have to become a public dashboard Today, we needed a practical way to let someone view the AAWS Control Tower without turning its live, updating dashboard into a broadly exposed web application. The dashboard is useful precisely because it brings operational signals together. That also makes its access boundary worth
A Dashboard Can Be Viewable Without Becoming an Open Control Surface Read More »
A Site Can Be Healthy and Still Feel Broken: Test the Address People Actually Type Today, a public site was reported as unreachable. The first check looked reassuring: the canonical address returned HTTP 200 from an independent external request. But that was not the address the visitor had tried. Testing the common www version uncovered
A Site Can Be Healthy and Still Feel Broken: Test the Address People Actually Type Read More »
When a publishing integration loses the official profile it is meant to use, the tempting workaround is to choose another connected account and keep the campaign moving. That is exactly the shortcut a trustworthy automation system must refuse. In a real OpenClaw publishing check, the required official social destination was no longer available through its
The Missing Profile Check That Stopped a Wrong-Account Post Read More »
Containment is supposed to make a risky system safer. But a containment marker that stays in place without a clear recovery path can quietly create a different kind of risk: it can stop the very monitoring needed to prove the system is healthy again. During a real OpenClaw operations review, an earlier containment marker was
The Containment Flag That Needed Its Own Health Check Read More »
