howardcha@google.com skills-eng@google.com Records whether a user saved a skill using the refined prompt, reverted it, or saved without refinement. This is tracked using undo/redo click counts at the time of save. This is a rough estimate. If the user manually edits the prompt after refinement without clicking undo, it is counted as using the refined prompt as it served as the base. michelleabreo@google.com skills-eng@google.com Records user actions (e.g., Save, Cancel, Refine) specifically within the Skills {DialogType} dialog. Invoked from {EntryPointAndState}. aashnas@google.com skills-eng@google.com Records the result status (success, network error, parse error, etc.) of a first-party skills download attempt. aashnas@google.com skills-eng@google.com The HTTP response code received when downloading skills. michelleabreo@google.com skills-eng@google.com Records executions of Skills, categorized by the skill source. michelleabreo@google.com skills-eng@google.com Records the terminal outcome (success or specific error state) of invoking a Skill via the Glic panel. This metric is used to calculate availability and reliability SLOs. Specific failure states include panel timeouts and requests for skills that no longer exist (e.g., deleted via sync). aashnas@google.com skills-eng@google.com Records errors encountered during skills management operations (e.g., service not ready, skill not found). aashnas@google.com skills-eng@google.com Records the status of a skills download request (success, already running, timeout). aashnas@google.com skills-eng@google.com Records user actions on the Skills {Page}, such as adding, editing, deleting, or searching skills. howardcha@google.com skills-eng@google.com Records the end-to-end latency (in milliseconds) of a Skills prompt refinement request, from initiation to result parsing. This metric is recorded for both successful model executions and asynchronous model failures, but not for synchronous local validation failures. michelleabreo@google.com skills-eng@google.com Records the terminal outcome of refining a Skill prompt using a server-side ML model. Tracks successful execution, model failures, protobuf parse errors, and cases where the model returns an empty suggestion list. Used to monitor feature health and model reliability. esalma@google.com austinzzr@google.com cbe-productivity-eng@google.com Records the source of a derived skill when it is successfully saved. Emitted in SkillsServiceImpl::SetAndSaveSkill when the skill is derived from an Enterprise or First Party skill. michelleabreo@google.com skills-eng@google.com Records the terminal outcome of saving a new Skill or updating an existing one. Failure states differentiate between database/sync errors, initialization states (service not ready), and cases where the UI context is lost (dialog closed) before the save operation completes. michelleabreo@google.com skills-eng@google.com Records the total number of skills a user possesses per profile. Used to measure adoption depth over time. This metric is recorded, potentially multiple times (once per profile) with every UMA log by SkillsMetricsProvider. michelleabreo@google.com skills-eng@google.com Records whether the postMessage handshake between the Skills host and guest Webview succeeded or timed out (after 5 seconds). michelleabreo@google.com skills-eng@google.com Records the duration of the {Stage} loading stage for the Skills Webview V2. michelleabreo@google.com skills-eng@google.com Records the total initialization latency of the Skills Webview V2 on startup, from the beginning of init() to handshake completion. michelleabreo@google.com skills-eng@google.com Records the time taken for the guest Webview to successfully save/update a skill over the network.