Skills & Agents
Every agent, A to Z
Every one of the 73 QA agents in the testland-qa marketplace, by name, with what each one does and the plugin that installs it.
A6
accessibility-code-criticHunts div onclick handlers, missing focus management, color-only cues, misused ARIA and missing labels, and reports each finding with its WCAG criterion and a fix.qa-accessibilityai-test-curatorFlags hallucinated APIs and imports, weak assertions such as toBeTruthy, redundancy with existing tests and missing setup or teardown, and refuses to mark tests ready while a high-confidence issue remains.qa-ai-assistedai-test-shallow-coverage-criticScores each public entry point for equivalence partitions, boundary values and error paths, and gives PASS, SHALLOW or N/A per axis with evidence.qa-ai-assistedapi-test-authorPicks Postman, REST Assured, Karate, Schemathesis or RESTler from the project's facts and writes the test in that tool's idiom.qa-api-testingasync-job-health-criticFlags missing retry limits, no dead-letter routing, unbounded backoff and non-idempotent handlers, with a findings table and a BLOCK or PASS verdict.qa-async-jobsautomation-harness-bootstrapperCovers folder layout, base fixtures, a page-object or screenplay base class, one example smoke test and the CI job.qa-web-e2e
B3
bdd-scenario-authorWrites the Gherkin feature file, detects the runner (Cucumber, Behave, Reqnroll, SpecFlow) and scaffolds step definitions, reusing existing steps first.qa-bddbug-report-from-recordingEmits a report in the bug-report-template schema: verbatim errors, repro steps from the recorded actions, environment from the trace, expected vs actual.qa-bug-reprobug-repro-builderFinds the smallest code unit that shows the failure, writes the failing test in the project's framework and runs it once to confirm it fails.qa-bug-repro
C7
cache-key-collision-detectorFinds cache keys missing tenant, user, locale or region scoping, missing Vary headers on CDN responses, and keys that hash differently across browser, CDN and Redis tiers.qa-cache-testingci-defect-filerSearches for an existing issue four ways, then files a new one or adds a recurrence comment.qa-defect-managementci-pipeline-health-criticFlags missing cancel-on-supersede, retries that hide flakiness, no test-report artifacts, no sharding on large suites and missing fail-fast or required gates, with a BLOCK or PASS verdict.qa-ci-integrationcode-quality-criticMerges and dedupes their findings, separates fix-now from ratchet-later, and refuses to mark the PR ready while it adds net new debt.qa-code-qualityconcurrency-criticLooks for unguarded shared mutable state, inconsistent lock ordering (deadlock risk), missing memory visibility and check-then-act races, with a findings table and a BLOCK or PASS verdict.qa-concurrencycontract-drift-investigatorDiffs Pact files, OpenAPI, GraphQL SDL or Protobuf against the last green baseline and says whether the provider, the schema, the consumer expectation or the test data changed, with file:line references.qa-contract-testingcontract-test-scaffolderEmits Pact consumer expectations, Schemathesis runners or buf breaking compatibility tests.qa-contract-testing
D6
daily-test-suite-aggregatorIngests the run artifacts and reports what passed, failed and changed across suites and environments.qa-test-reportingdata-anomaly-triagerClassifies the failure as drift, outlier, missing, referential or freshness, names the likely owning team and proposes the remediation step.qa-data-qualitydb-snapshot-restoreCaptures a baseline snapshot (Postgres template DB, mysqldump, Mongo dump), restores it before each test or suite, and wires BEGIN/ROLLBACK where the ORM allows.qa-test-environmentdefect-fix-verifierChecks that the build under test contains the fix commit, re-runs the repro, returns VERIFIED, NOT FIXED or BLOCKED with evidence and moves the ticket to Verified or Reopened.qa-defect-managementdefect-pipeline-runnerClusters bugs by root cause, writes the Pareto trend with week-over-week deltas and escape rate, classifies each production escape as a test, process or tooling gap, and assembles one report.qa-bug-reprodesktop-test-authorDetects the driver (FlaUI, WinAppDriver, Playwright for Electron, QtTest, XCUITest, AT-SPI), scaffolds the project if needed and writes one test per user flow.qa-desktop
E4
e2e-flake-bisectorReruns it under varied isolation, order, viewport, network and parallelism to find the axis that reproduces it, names the shared resource when workers collide and reports a reproduction rate per axis.qa-flake-triagee2e-selector-quality-criticFlags CSS class, nth-child, XPath and position selectors and proposes getByRole, getByLabel and other accessibility-first locators.qa-test-reviewembedded-test-authorDetects Ceedling, CMake with GoogleTest or a bare Makefile, scaffolds the harness if needed, and writes one test file per spec in Unity, GoogleTest or Ceedling with CMock mocks.qa-embeddedextension-test-authorDetects Manifest V2 or V3 and the target browser and writes one Playwright spec per behavior.qa-browser-extension
F3
flag-coverage-gap-detectorFinds flag evaluation call sites whose off path, fallthrough path or non-default variants no test exercises, and ranks the untested branches.qa-feature-flagsframework-architecture-auditorAudits page-object consistency, base-class depth, fixture coupling, helper sprawl, naming and wait conventions, convention drift, CI health and dead helpers across files.qa-test-reviewfuzz-target-authorPicks libFuzzer, AFL++, cargo-fuzz, Go fuzzing, Atheris or Jazzer by language and writes the harness, seed corpus, dictionary, build command and CI snippet.qa-fuzz-testing
G4
game-test-scenario-authorDetects the engine, scaffolds the test structure if needed, and writes one scenario per behavior that drives inputs and asserts state transitions.qa-gamegherkin-style-reviewerFlags imperative UI steps, technical leakage (selectors, URLs, table names), long And chains, repeated Givens that belong in Background, and Then steps with no observable outcome.qa-bddgolden-file-managerAdds missing baselines, updates baselines only when the diff matches the PR's stated intent, prunes orphaned ones and applies the sanitization rules.qa-test-datagrpc-service-reviewerChecks for missing status-code assertions, deadline tests, buf lint and buf breaking in CI, untested streaming RPCs and a client mock harness, with a ranked findings table and a BLOCK or PASS verdict.qa-grpc
M2
migration-blast-radius-reviewerClassifies each operation as additive, breaking, data-loss or locking, estimates downtime on large tables, finds downstream consumers and flags missing rollback, unsafe defaults and missing CONCURRENTLY.qa-db-migrationsmobile-test-authorDetects XCUITest, Espresso, Detox, Flutter, Appium or Maestro from the project, scaffolds the harness if needed and writes one test per behavior in that driver's idiom.qa-mobile
P5
payment-flow-criticChecks for missing idempotency keys on charges, unverified webhook signatures, card data in logs, unhandled 3DS requires_action states and double-charge races, with severity, fixes and a BLOCK or PASS verdict.qa-paymentplaywright-codegen-reviewerRefactors it into idiomatic tests: page objects or custom commands for repeated steps, role-based selectors, AAA structure, and retry-aware waits instead of fixed sleeps.qa-web-e2eprompt-eval-reviewerFlags too few cases, one provider only, missing model-graded or semantic checks, no baseline diff, no cost or latency cap, unpinned models and no adversarial cases.qa-llm-evaluationproperty-based-test-authorPicks fast-check, Hypothesis, jqwik or proptest by language and writes the property with generators the framework can shrink.qa-property-basedpwa-test-authorDetects the surface from the manifest and service worker registration and writes one Playwright spec per behavior.qa-pwa
Q2
qa-managerReads CI history, the defect tracker and flake quarantine, computes pass-rate trend, escaped defects and flake debt, and writes a one-page red, amber or green status per area.qa-team-managementquality-coachReads the team's Definition of Done and tags each line met, not met or unverifiable against the PR's evidence, so unfinished work does not ship as done.qa-process
R7
realtime-protocol-reviewerFlags missing reconnect with backoff, unhandled close codes, missing webhook signature checks, wrong MQTT QoS or Clean Start and missing heartbeats, with a ranked findings table and a BLOCK or PASS verdict.qa-realtime-protocolsregression-bisectorDrives git bisect (including a perf-threshold mode) to the introducing commit and reports the culprit.qa-flake-triageregression-suite-curatorFinds duplicates, tautologies, dead-signal and orphan tests with evidence and proposes keep, fold or delete through a four-condition delete gate.qa-test-impact-analysisrelease-engineerRecords each step and writes a release report, and always pauses for human approval at decision points.qa-processrelease-quality-report-agentCombines the test-run narrative, the per-file coverage diff and risk-ranked coverage gaps into one report with a GO, NO-GO or CONDITIONAL verdict.qa-test-reportingrelease-readiness-checkerRuns every gate in release-readiness.yml (smoke, coverage, open critical bugs, migration dry run, threat model and more) and returns one go or no-go verdict with evidence per gate.qa-processrelevance-regression-reviewerRuns the judgment list before and after, reports NDCG, MRR and Recall@k deltas per query, flags regressions and refuses to approve a net relevance drop.qa-search-relevance
S6
saga-criticFlags missing or non-idempotent compensations, dual writes without a transactional outbox and steps with no retry policy, with a BLOCK or PASS verdict.qa-saga-cqrssample-ratio-mismatch-detectorRuns a chi-square test on observed vs expected allocation and, when there is a sample ratio mismatch, lists the likely causes such as logging bugs, bot filtering, redirects or randomization bugs.qa-feature-flagsschema-diff-reviewerSeparates breaking from additive changes, finds new or changed columns with no data tests and lists downstream consumers affected.qa-data-qualitysecurity-finding-triagerNormalizes and dedupes findings, enforces waivers with expiry and approver, adds EPSS and CISA KEV, and posts one PR comment with a BLOCK or PASS per domain.qa-security-scanningserverless-cold-start-criticFlags clients created inside the handler, heavy top-level imports, no /tmp reuse, missing SnapStart on JVM runtimes and packages over the platform size limits, with a BLOCK or PASS verdict.qa-serverlessstale-flag-detectorFinds flags at 100% rollout, untouched kill switches, shipped experiments and orphan flags, ranks them for removal and gives a safe removal runbook for each.qa-feature-flags
T12
tcm-migration-agentMaps fields, exports via API, transforms, dry-runs the import, then runs it and writes a field map and per-case log.qa-test-managementtenant-leak-criticTraces tenant_id from each entry point to queries, caches, logs and messages and flags missing tenant filters, untrusted tenant input, unprefixed cache keys and RLS gaps.qa-multi-tenancyterraform-plan-reviewerReads the plan JSON and flags resource destruction, broader IAM, public exposure, disabled encryption, drift and risky combinations such as a database destroy plus recreate, with severity and a fix for each.qa-iactest-case-quality-criticChecks required fields, one action per step, title and expected-result quality, requirement links, orphans and coverage.qa-test-managementtest-code-criticFlags structure problems (AAA, one behavior per test, naming, magic numbers, fixture coupling), vague assertions and over-mocking, with file:line evidence.qa-test-reviewtest-data-setup-agentDetects the stack, generates per-test fixture factories and a reproducible E2E seed dataset, and wires both into the test bootstrap.qa-test-datatest-environment-bootstrapperStarts containers with Testcontainers or Docker Compose, runs migrations, seeds a baseline database, installs an in-memory OpenFeature provider and emits Playwright fixtures for auth, database and flags.qa-test-environmenttest-script-quality-criticFlags vague preconditions, bundled steps, missing or ambiguous expected results, non-self-contained data and UI mechanics, and gives PASS or BLOCK per script.qa-manual-testingtest-suite-health-auditorInventories files, classifies tiers, checks the pyramid ratio, flake rate and ROI per layer and selector and assertion quality, and returns Healthy, Needs pruning, Needs refactor or Cannot assess with top fixes.qa-test-reviewtime-handling-criticFlags naive now() calls, local-time arithmetic, DST-unsafe construction, naive date storage and missing zone-aware types, with severity and a BLOCK or PASS verdict.qa-timetoken-storage-security-criticFlags tokens in localStorage or sessionStorage, cookies missing httpOnly, Secure or SameSite, tokens in logs or URLs, and missing rotation or expiry, with a BLOCK or PASS verdict.qa-auth-flowstrace-coverage-reviewerFinds untraced database calls, HTTP calls, queue operations and cron jobs, spans missing semantic-convention attributes, hand-rolled spans that auto-instrumentation covers and cardinality risks.qa-distributed-tracing
V3
vacuous-property-criticFlags preconditions that discard most inputs, properties with no real assertion, generators too narrow for the invariant and assertions that restate the code.qa-property-basedvisual-baseline-curatorProposes a coverage-optimal baseline set for Storybook or app pages and writes the Chromatic, Percy, Playwright or Storybook test-runner config.qa-visual-regressionvisual-diff-classifierClassifies each diff as intentional, incidental or regression with a reason and action, and on large builds clusters diffs by component or route so reviewers check the right screenshots.qa-visual-regression