← Back to portfolio

Case study 01 / Systems & IoT

Taking control.
Keeping it usable.

An LG webOS home lab combining community Homebrew, key-based administration and a replacement home screen—with restart checks and a route back.

PlatformLG webOS 10.3.2
Completed21 October 2026
ScopePersonal, owner-controlled device
MethodPractical home lab

01 / The brief

A useful device, with more control.

The aim was to customise my television and make it maintainable, while preserving the functions it was bought for: games, streaming, sound and a working remote.

The test device was my LG 50UT8050PSB, running firmware 33.31.75 and webOS 10.3.2. The project used existing community tools to obtain administrator access, install Homebrew Channel and integrate Glasshouse as the home interface.

Successful installation was only one requirement. The changes also needed to survive a restart, retain a working stock-home fallback and leave the household’s normal networking restored.

Bounded scope

All device work took place on my own television and local network. This is a documented learning project using existing research and software. It is not an original exploit discovery or a comprehensive security assessment.

02 / The system

Separate everyday use from maintenance.

Conceptual home-lab architecture showing the TV, maintenance workstation and normal entertainment functions. Open diagram at full size.
Conceptual architecture. Open diagram at full size ↗

A temporary direct Ethernet connection supported installation and diagnosis. After the setup, the television returned to normal Wi-Fi use, and its wired configuration was restored to automatic addressing.

Administrator access used SSH with a key and a pinned host identity. Telnet was disabled. Glasshouse’s management listener was configured for the television’s loopback interface, with maintenance access handled separately.

The custom home interface sat alongside a tested LG Home fallback. That distinction mattered: changing the home experience did not require discarding the original interface.

03 / Decisions & troubleshooting

Where the learning happened.

01

Distinguish reachability from privilege.

The first launch route, adoverlayex, did not load the test page. A different route, voiceweb, did. That established that the television could fetch and display the page; it did not establish root access.

The installer path was adapted and actual administrator access was checked separately. Treating those as different milestones avoided declaring success from an intermediate result.

02

Make administration explicit.

The Homebrew package’s potential password-free Telnet access was replaced with the intended management arrangement: key-based SSH and disabled Telnet. The resulting startup behaviour was checked after a restart.

These were specific configuration checks. They do not prove that every service or vulnerability on the television was assessed.

03

Integrate startup with a recovery path.

Glasshouse was added to the existing startup arrangement. Prior configuration was retained, automatic launch was tested after restarting, and switching back to LG Home was exercised.

The wallpaper initially failed to appear. That became a separate configuration problem to resolve, followed by a visual check, rather than being hidden behind an otherwise successful install.

04

Close the installation window properly.

Temporary PC networking, firewall exceptions and the installer listener were removed or restored. The TV’s wired settings were returned to DHCP with manual DNS entries cleared.

Everyday testing then covered PS5 use and streaming applications. A working custom launcher mattered only if the television remained comfortable to use.

04 / Verification

What the evidence supports.

The following describes the completed setup and recorded checks from the project. It is a historical result, not a live status monitor.

On a small screen, swipe the table to read each result.

LG home-lab verification outcomes
CheckEvidenceOutcome
Homebrew and administrator accessHomebrew Channel 0.7.3 installed; root SSH verified.Verified
Management accessKey-based SSH; Telnet disabled.Verified
Restart persistenceStartup and management access checked after restart.Verified
Glasshouse integrationVersion 0.118.0 active and started automatically.Verified
RecoveryLG Home fallback and disable/re-enable tested.Verified
Temporary network changesPC restoration recorded; TV wired configuration returned to automatic.Verified
Everyday useOwner-confirmed PS5 picture/sound, appearance and streaming use.User-tested
PS5 Remote Play extensionRegistration and configuration transfer completed; playback and Bluetooth controller tests outstanding.In progress
Limits of the result

No measured speed improvement, complete telemetry removal, comprehensive debloating or independent penetration test is claimed. A known platform limitation also prevented one Glasshouse developer-inspector filtering function; the lab does not claim all services were filtered.

05 / What I learned

A successful change has a wider boundary.

Through this project, I developed a more concrete understanding of addressing and reachability, Linux startup behaviour, authenticated remote administration and configuration recovery.

The most useful lesson was to name each milestone precisely. Loading a page, launching a service, obtaining administrator access and retaining it after restart are different findings. Each needs evidence appropriate to the claim.

The project also connected technical choices to normal use. A recovery path, a working remote and reliable HDMI playback were part of the outcome, not housekeeping after the interesting work.

For documentation, the same principle applies: distinguish what was observed, what it demonstrates and what remains unknown. That approach carries directly into the way I write about alerts, application security and troubleshooting.

06 / What remains

A separate experiment: PS5 Remote Play.

Chiaki was installed as an extension to the completed Homebrew setup. Desktop registration succeeded and the resulting configuration was transferred to the television with exact readback verification.

The released TV importer had a known byte-array handling problem and logged pairing material. A format converter was prepared to decode the desktop export correctly, validate expected byte lengths and create the TV configuration directly. Synthetic tests exercised escaping and malformed inputs. Sanitised source and tests are included in the source download.

At publication, the LG app had launched, but actual streaming video, audio and Bluetooth controller behaviour had not been verified. It remains in progress. Those tests are the next acceptance step; installation alone is not a completed Remote Play setup.

07 / Credits & method

Built with community knowledge.

I owned and operated the hardware, chose the project goals and appearance, supplied observations, and carried out physical restarts and playback tests.

The published scripts support configuration conversion and verification. The exploit and the applications are the work of their upstream developers, credited below. This case study documents their integration in my home lab.

Private keys, pairing exports, account identifiers and raw device logs are excluded from this publication. The source download contains selected source, synthetic tests and a sanitised verification record.

Continue the conversation

Questions about
the project?

I’m happy to discuss the decisions, the learning and the next tests.