docs: show native HTTPS HAR recording lifecycle

[MagMueller]

Clarifies the existing Python Browser HAR recorder: it captures HTTPS, writes the file on session stop, and omit mode removes request/response bodies but does not redact URLs or headers. Adds one local-browser example with a unique output directory and finally/kill cleanup.

Verified on Browser Use main 843819cb using owned loopback HTTP/TLS fixtures. HTTP rendered successfully but produced an empty HAR. HTTPS produced document and POST entries after shutdown. omit and embed behavior, retained synthetic headers/query, and process cleanup were exercised. The documented function also ran with only fixture URL/certificate/output-path adaptations. No model, Cloud API, or public website was used. This does not claim a complete all-tabs or worker archive.

The recipe is unchanged from the previously tested candidate. MDX parsing, Mint links, docs JSON and diff checks passed. One authored docs file only; generated exports removed upstream are not restored. No runtime or dependency change.

---

Summary by cubic

Clarifies the existing Python Browser HAR recorder docs: it captures HTTPS, writes the HAR when the browser session stops, and record_har_content='omit' removes request/response bodies without redacting URLs, query strings, or headers.

• Adds a local-browser example using a fresh output directory and finally/kill cleanup.
• Notes that HTTP-only visits can produce empty HAR files and that recording does not recover earlier traffic.

^{Written for commit 1671b31. Summary will update on new commits.}

[cubic-dev-ai]

No issues found across 1 file

_{Re-trigger cubic}