autonomous-software-testing

Plugin autonomous-software-testing from marketplace autonomous-testing-marketplace in Purushotham-Prajapati-24/qa-skills. Commands read from the repo's manifest — 1 month ago.

Skills (21)

• accessibility-testing

  Test a user interface for accessibility — automated WCAG scanning, keyboard navigation, focus management, screen reader semantics, colour contrast, form labelling, and the manual checks automation cannot perform. Use when a web UI changed, when accessibility conformance is required, or when asked to

• ai-testing

  Test LLM-backed features, RAG pipelines and agent tool-calling — golden sets, rubric grading, retrieval quality, hallucination and grounding checks, prompt injection resistance, tool-call correctness, cost and latency, and determinism strategies for a non-deterministic system. Use when the repositor

• api-testing

  Test HTTP, GraphQL and gRPC APIs — status codes, response shapes, schema conformance, validation, error handling, authentication and authorisation, pagination, idempotency, rate limiting and backwards compatibility. Use when a change affects an API, when testing endpoints, validating a contract, or

• browser-testing

  Choose and run the right browser testing method — agent-driven exploration through a browser MCP, a deterministic committed script, the repository's existing browser suite, or explore-then-automate. Use when testing a web UI, exploring an application for bugs, deciding between Playwright MCP and Pla

• change-intelligence

  Work out exactly what changed in a pull request, commit range, branch or described feature, which modules and user-facing behaviours it affects, which existing tests cover it, and where the coverage gaps are. Use when testing a PR or commit, deciding which tests to re-run, scoping regression testing

• compatibility-testing

  Test across browsers, viewports, devices, locales and runtime versions — cross-browser behaviour, responsive layout, mobile viewports, localisation and internationalisation, and dependency or runtime version compatibility. Use when a support matrix exists, when a UI must work on more than one browse

• database-testing

  Test data layer behaviour — queries, constraints, transactions, data integrity, schema migrations forward and backward, seed and fixture correctness, and the data-safety implications of a change. Use when a change touches the database, adds a migration, alters a schema, or when data correctness is i

• defect-reporting

  Turn a test failure or observation into an actionable defect report — reproduction steps, expected versus actual, severity, impact, evidence, confidence, and a GitHub or Jira issue that a developer can act on without asking questions. Use when a real defect is found, when writing up a bug, or when d

• e2e-testing

  Test complete user journeys against a running application — critical paths, cross-page flows, real integration, authentication states and data setup and teardown. Use when a change affects a journey a user completes end to end, when validating a critical business flow, or when pre-release confidence

• integration-testing

  Test how components work together — database access, queues, caches, external service boundaries, contracts, concurrency, resilience and fault injection, observability, infrastructure and CI/CD configuration, deployment and rollback. Use when a change crosses a module or system boundary, when integr

• performance-testing

  Measure and test performance — latency, throughput, load, stress, spike and endurance testing, plus frontend performance — against stated requirements and a measured baseline. Use when a change affects a performance-sensitive path, when latency or throughput requirements exist, or when performance t

• regression-testing

  Re-run existing tests intelligently against a change — selecting which suites matter, running smoke and sanity checks, detecting flaky tests from run history, distinguishing a new failure from a pre-existing one, and turning fixed defects into permanent regression coverage. Use when validating that

• repository-intelligence

  Build a structured, evidence-backed profile of a repository — languages, frameworks, architecture, APIs, databases, queues, caches, auth, external integrations, AI/LLM systems, existing test frameworks and coverage, CI/CD, containers, infrastructure and observability. Use before planning any testing

• requirement-analysis

  Find and read the requirements that testing should validate — Jira tickets, GitHub issues, acceptance criteria, specs, README claims and the code itself — assess which are relevant to the change at hand, extract testable criteria, and flag ambiguity instead of inventing expected behaviour. Use when

• risk-analysis

  Score the risk of a change or component from evidence — business criticality, security sensitivity, user impact, data sensitivity, change magnitude and frequency, dependency and integration complexity, historical failures, coverage deficit, flakiness, deployment exposure, irreversibility and externa

• security-testing

  Test authentication, authorisation, session management, input validation and dependency security within an authorised scope — broken object-level authorisation, privilege escalation, session handling, injection, secrets exposure and vulnerable dependencies. Use when a change touches auth, permission

• test-reporting

  Produce the testing report and maintain the versioned testing log — executive summary, evidence index, findings, blocked, deferred and interrupted work, coverage gaps, traceability, evaluation metrics and an integrity self-audit. Use when testing is finished or paused, when the user asks what was te

• test-strategy

  Turn a repository profile, change analysis, requirements and risk assessment into a concrete test plan — scope, out-of-scope, strategy, depth, prioritised scenarios, test data, environments, tooling, expected evidence, success and exit criteria, fallbacks, deferred work and open questions. Use when

• testing-orchestrator

  Plan and run software testing for a repository, a pull request, a commit, a feature or a ticket. Discovers the codebase, works out which kinds of testing actually apply, scores risk, chooses tools, executes tests, collects evidence, records findings, and reports honestly on what passed, failed, was

• ui-testing

  Test user interface behaviour at the component and page level — rendering, state transitions, form validation, loading and error states, empty states, keyboard interaction and visual regression. Use when a UI change needs verifying below full end-to-end level, or when adding component or visual test

• unit-testing

  Write, run and assess unit and component tests — the repository's existing framework and conventions, meaningful assertions, boundary and negative cases, test doubles, determinism and coverage interpretation. Use when unit-level testing is applicable, when adding tests for changed functions or compo

Subagents (4)

• browser-explorer

  Drives a browser to explore a running application and report what it finds — user journeys, UI states, console errors and failed requests. Use for exploratory browser testing when the interaction transcript would flood the main context. Never use it as a test gate.

• evidence-auditor

  Adversarially audits testing claims against the evidence on disk. Use before delivering a report, or whenever a testing claim needs independent checking. Returns only the claims that are not supported.

• repository-analyst

  Read-only codebase surveyor. Use when profiling a repository would flood the main context with file listings and source excerpts — large repositories, monorepos, or unfamiliar stacks. Returns a structured repository profile, not raw output.

• test-author

  Writes automated tests that match a repository's existing conventions. Use when adding several test files at once, or when the existing test suite must be read in depth before writing. Returns the files written and why each assertion exists.

MCP servers (1)

• playwright

  runs: npx @playwright/mcp@latest --isolated --headless --output-dir ${CLAUDE_PLUGIN_DATA}/browser-artifacts

Hooks (4)

• SessionStart
• UserPromptSubmit
• Stop
• PreToolUse

Similar plugins