fix: include captured browser failure screenshot

[MagMueller]

The browser failure action captures and base64-encodes a screenshot, but its format string returns the literal <TODO:insert the screenshot>. The supplied screenshot_base64 argument is unused.

Replace that placeholder with {screenshot_base64}. This changes one line in one file; the user-journey worker still invokes this agent.

Validation against current main af6d2c1:

• Executed the actual format-call AST with PNG_BYTES: main returns the literal TODO; the patch includes PNG_BYTES.
• Python compilation and git diff --check pass.
• Ruff F522 disappears. Four pre-existing findings remain unchanged: I001 import order, F401 unused FailureScreenshot, and two ASYNC230 blocking file opens.
• Ruff format reports the same pre-existing indentation change on both versions.

No paid model call or full browser journey was run. No unrelated cleanup is included.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.