Source: the obtained 3.7.0 release and versioned project documentation.
Begin with the failing boundary
A browser catalog and a local runtime can fail independently. During restoration, the old mirror API returned HTTP 200 with list: null and total: 0; the newer public store API returned records. That observation identifies an empty old service response. It does not show that the application ecosystem contains no software or that a visitor’s machine is broken.
Four questions to answer
Does the catalog request return usable records, an explicit empty result, a timeout or an error?
Does the local command exist, and which version is installed?
Can the local package manager read repository configuration?
Did installation fail, or did a successfully installed application fail to open?
ll-cli --version
ll-cli --help
ll-cli --json repo show
The last command is documented in the 3.7.0 store’s environment-detection design. If your version does not support it, stop and use its matching reference. Record command output and the exit code rather than treating an empty UI as a diagnosis.
Worked incident: repository read fails
The source describes a case where ll-cli is installed but repository inspection fails because the package-manager service is unavailable. That differs from having no repository configured. The client’s recovery requires an explicit user action and local authorization; this website does not restart services or offer a universal root repair script.
A report another person can act on
Store version:
Distribution and CPU:
ll-cli version:
Task: browse / inspect repository / install / launch
Application ID and repository:
Exact failed command or URL:
Exit code and original error:
Expected result:
Actual result:
Reproduction steps:
Sensitive information removed:
The source distinguishes an original machine error from a translated display message. Keep the original error for reproduction, but review logs for tokens, personal paths and identifiers before sharing them. A recovery process finishing successfully still needs a new attempt at the original task.
An installed application that does not open needs a launch report, not a claim that the catalog is empty. Keep the failed boundary visible. The following filled structure is synthetic; it records no incident on a real machine.
First record the distribution, CPU architecture and exact local client version. These commands inspect information; they do not install a package or repair a service. Do not run them on a machine you do not administer.
ll-cli --version
ll-cli --help
ll-cli list
Find the application in the installed list. Copy its exact identifier/version/channel/architecture as displayed; do not substitute a similarly named catalog entry. If it is absent, stop: the question is installation or the selected environment, not a reproduced launch failure.
Launch it using your usual application menu and record the exact action, visible result and original error. If no error appears, write “no error shown”. Do not invent an exit code from a graphical launcher. An optional command-line reproduction must use the syntax printed by your installed ll-cli --help.
Compare scope: one application, several applications, or not yet checked. Avoid uninstalling, changing repositories or restarting services merely to gather evidence. Such changes can destroy the useful before-state.
Review the report locally. Replace usernames, home paths, private URLs, access tokens and unrelated application names before sharing. Preserve the wording of the actual error and mark every redaction. Attach only the smallest relevant excerpt; never upload an entire home directory or unrestricted system log.
Example only — not a measured failure
Task: launch an already installed application
Distribution / CPU: [fill from your machine]
ll-cli version: [paste exact output]
Application ref: [copy matching entry from ll-cli list]
Before: application appears in installed list
Action: launch from the usual application menu
Expected: one application window
Observed: no window; error text [paste exact text]
Reproduction: 1. launch 2. observe 3. repeat once if safe
Scope: one application / all applications / unknown
Removed: username, home path, tokens, private URLs
Unknown: cause; no installation or service repair attempted
The checked upstream CLI source is pinned to commit 80f57cfe (13 November 2026); its observed latest release is 1.14.3, published 24 September 2026. These identifiers describe the references, not your installed version. The three reading commands are present in the source; no Linyaps installation, launch, fix or Linux benchmark was executed here. A report is complete when another person can distinguish the observed symptom from the still-unknown cause.
Keep observation separate from cause. Diagram: Linglong, CC BY 4.0.