.claude/agents/

The 28 K0lmenIA agents

Each agent is a specialist with its own input, process and output. Claude Code picks the right one for your request, or you can name it directly.

Analysis and design

8 agents

analista-historias

Researches the story's context and spots ambiguities, gaps, risks and missing information.

output/analisis-historias/

estratega-pruebas

HTML test plan: scope, risks, test types, environment and criteria.

output/planes-de-prueba/

generador-casos-manuales

Cases in Excel and Markdown plus a coverage report with questions for the PO.

output/casos-de-prueba/manuales/

generador-casos-bdd

Gherkin scenarios (English keywords, Spanish content) plus coverage.

output/casos-de-prueba/bdd/

generador-casos-api

Positive, negative, schema and authorization API cases.

output/casos-api/

generador-datos-prueba

Valid, invalid and edge-case data in Markdown or CSV.

output/datos-de-prueba/

selector-casos-de-prueba

Crosses a diff, PR or release notes with your cases and builds a risk-based selection, with the npm test command ready.

output/seleccion-casos/

generador-reportes-bug

Professional bug report with justified severity and priority.

output/reportes-bug/

Live execution

10 agents · uses tokens on every run

ejecutor-e2e

Runs cases in a real browser with Playwright MCP, with evidence.

output/ejecuciones/

ejecutor-api

Runs a Postman collection with Newman.

output/ejecuciones/

explorador-web

Exploratory testing: broken links and images, JavaScript errors and responsive checks, with screenshots and a report.

output/exploratorias/

analista-accesibilidad

WCAG 2.2 audit with axe-core, keyboard and reflow, plus manual review.

output/accesibilidad/

auditor-rendimiento-web

Lighthouse on mobile and desktop: Core Web Vitals and opportunities with estimated savings.

output/rendimiento/

pixel-perfect

Visual regression against a baseline or the design: what moved and by how many pixels.

output/pixel-perfect/

cross-browser

Chromium, Firefox and WebKit: isolates what fails in only one browser.

output/cross-browser/

generador-reporte-html

HTML dashboard of a run's results.

output/ejecuciones/

generador-reporte-cierre

Round closing report with a go/no-go recommendation.

output/informes-cierre/

analista-cobertura

Traceability matrix from story to result, and what's left to test.

output/cobertura/

Automation with k0lmena

8 agents · then runs with zero tokens

web-mapper

Walks the web app following your cases and writes .feature files, steps and locators.

k0lmena/web/

api-mapper

Reads Postman or Swagger/OpenAPI, verifies the endpoints and writes the .feature files.

k0lmena/api/

mobile-mapper

Walks the app with Appium and writes .feature files, steps and locators.

k0lmena/mobile/

performance-mapper

Load, stress, soak and spike tests with k6, Artillery or JMeter; asks for load and thresholds.

k0lmena/performance/

verificador-datos

Checks data in PostgreSQL, MySQL, SQL Server or MongoDB and adds database checks to tests.

steps .feature

analista-fallos

When npm test fails, separates real bugs from broken, flaky, data or environment issues.

output/fallos/

reparador-automatizacion

Diagnoses broken locators or waits using the saved DOM, applies the fix with a backup and verifies it.

output/reparaciones/

analista-seguridad

Passive web security scan with OWASP ZAP, with findings and how to fix them.

output/seguridad/

Test management

2 agents

gestor-pruebas

Folders, cases, story links and cycles in Xray, QMetry (QTM4J), AIO Tests or Azure DevOps Test Plans.

output/gestion/

publicador-resultados

Uploads npm test results to the cycle: status, error and evidence.

Xray · QMetry · AIO

Need an agent that isn't here?

K0lmenIA can be extended with new agents, skills and connectors. We can build the ones your process needs.

Ask about a custom agent