evals-write-spec
Write LLM evaluation spec files with datasets, tasks, and evaluators using the @kbn/evals Playwright fixture. Use when authoring new eval specs, adding datasets or evaluators, or debugging evaluation test failures.
Browse reusable Agent Skills, each with a clear purpose and practical guidance.
Write LLM evaluation spec files with datasets, tasks, and evaluators using the @kbn/evals Playwright fixture. Use when authoring new eval specs, adding datasets or evaluators, or debugging evaluation test failures.
Fix API documentation issues in a Kibana plugin or package. Use when asked to fix, improve, or add JSDoc/API documentation for a Kibana plugin or package, or when check_package_docs validation fails.
Security Solution specific. Analyzes flaky, failing, or skipped Security Solution Cypress tests to determine root cause and recommend fixes. Use when: (1) a user shares a flaky or skipped Cypress test, (2) asked to fix a test that intermittently fails, (3) asked to unskip a test, (4) triaging a test stability GitHub issue, (5) asked "why is this test flaky/failing", (6) asked to analyze test failures in ESS/Serverless/MKI environments, (7) asked whether to fix, delete, or migrate a broken Cypress test.
Investigate Scout and FTR flaky test failures in Kibana. Use when triaging a failed-test issue, a Buildkite-reported failure, a test path that has been failing intermittently, or any time the user asks to look at a flaky test, deflake a test, or stabilize a test.
Fleet-specific. Analyzes failing, flaky, or skipped Fleet tests to determine root cause and recommend the right action. Three modes: (1) triage a single GitHub issue, (2) audit the full Team:Fleet failed-test backlog, (3) fix a cluster of related issues together. Use when: (1) a user shares a failing or skipped Fleet test, (2) asked to triage or fix a flaky test, (3) asked to audit or clear the Fleet failed-test backlog, (4) asked to find the shared cause across a group of related test failures.
Deep reference for the Kibana Functional Test Runner (FTR). Use when reading, analyzing, debugging, or reviewing FTR tests, including config anatomy, services, page objects, loadTestFile patterns, data loading, tags, CI wiring, and common FTR idioms.
Use when an agent needs high-fidelity Elastic Security development data while working in Kibana: endpoint events, endpoint alerts, Security detection alerts, Attack Discoveries, or generated cases. This is for local engineering development, testing, and debugging only, not cloud deployments or customer-facing demos.
Create a new GitHub issue (feature request or bug report) by gathering an unstructured description from the user, classifying it, filling out the appropriate Kibana template, interviewing the user to improve any weak sections, and filing the issue via the GitHub CLI.
Fetch a GitHub issue by number or URL and reformat it according to the Kibana bug report or feature request template. Classifies the issue type automatically, then rewrites the body and updates it via the GitHub CLI.
Add or improve OpenAPI descriptions, examples, and code samples for a Kibana API area. Use when a plugin or package needs better OAS documentation — missing descriptions on schema fields, missing route summaries, missing examples, or missing code samples.