From wicked-testing
Tier-1 orchestrator for test planning. Covers test strategy, risk, testability review, and requirements quality. Dispatches specialist agents based on what the target needs. Use when: "what should I test", "test strategy", "test plan", "risk matrix", "is this testable", "are these requirements testable", "coverage strategy", "shift-left testing".
How this skill is triggered — by the user, by Claude, or both
Slash command
/wicked-testing:planThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
One skill for everything before tests get written. Figures out what to test,
One skill for everything before tests get written. Figures out what to test, what can go wrong, and whether the design lets you test at all.
Read the target first, then route:
| Target | Dispatch |
|---|---|
| Acceptance criteria / clarify doc | wicked-testing:requirements-quality-analyst |
| Design doc / architecture sketch | wicked-testing:testability-reviewer |
| Feature description / user story | wicked-testing:test-strategist |
| Known-risky change (security, data, perf) | wicked-testing:risk-assessor |
| "Test everything" / broad review | All four in parallel; merge findings |
When multiple apply, dispatch in parallel. Merge results in the reply — no unrelated raw outputs dumped in.
Task(
subagent_type="wicked-testing:test-strategist",
prompt="""Generate a comprehensive test strategy for the target below.
## Target
{file path, directory, or feature description}
## Instructions
1. Classify the change type (UI, API, both, data, config).
2. Analyze the surface area (public APIs, functions, endpoints).
3. Generate positive + negative scenario pairs for every feature.
4. Identify risk areas and confidence level.
5. Flag any specification gaps discovered.
**MANDATORY**: Every scenario must have BOTH positive AND negative counterpart.
Return findings in the standard test-strategist format."""
)
Swap subagent_type to the matching agent from the table above. For the
"test everything" path, dispatch all four in parallel (one Task(...) call
per agent in the same turn) and merge the returned findings.
For domain-specific planning signals, dispatch a specialist and fold its output into the strategy document. These don't render verdicts — they add risk+scenario coverage where the generalist agents would miss signal:
| Trigger (anything in the target that matches) | Specialist |
|---|---|
| React/Vue/Svelte component under test | wicked-testing:ui-component-test-engineer |
| API / service boundary (REST, gRPC, GraphQL) | wicked-testing:integration-test-engineer |
| Database migration or schema change | wicked-testing:data-quality-tester |
| Performance-sensitive path (heavy compute, I/O) | wicked-testing:load-performance-engineer |
| Multi-step user journey | wicked-testing:e2e-orchestrator |
| UI with visual regressions risk (CSS, theming) | wicked-testing:visual-regression-engineer |
| User-facing surface (WCAG 2.1 AA relevance) | wicked-testing:a11y-test-engineer |
| Parser / serializer / round-trip / invariants | wicked-testing:fuzz-property-engineer |
| Translated / RTL / pluralization-sensitive copy | wicked-testing:localization-test-engineer |
| Service with logs / metrics / traces / PII-in-signals | wicked-testing:observability-test-engineer |
| Test-suite effectiveness evaluation (kill rate) | wicked-testing:mutation-test-engineer |
| Failure-mode / resilience planning | wicked-testing:chaos-test-engineer |
| Application-security scope (SAST/DAST/authz) | wicked-testing:security-test-engineer |
| LLM / AI feature in the codebase | wicked-testing:ai-feature-test-engineer |
| Terraform / helm / k8s / IaC / policy-as-code | wicked-testing:iac-test-engineer |
| Regulated industry (SOC2 / HIPAA / GDPR / PCI) | wicked-testing:compliance-test-engineer |
| Test suite with snapshots (jest / syrupy / cassettes) | wicked-testing:snapshot-hygiene-auditor |
| Test-suite quality itself (smells, dead tests) | wicked-testing:test-code-quality-auditor |
| Prod incident -> regression scenario synthesis | wicked-testing:incident-to-scenario-synthesizer |
Tier-2 names are internal — see docs/NAMESPACE.md. Consumers (wicked-garden) depend only on Tier-1 names.
Emits wicked.teststrategy.authored on the bus when present.
docs/INTEGRATION.mddocs/NAMESPACE.mdagents/test-strategist.md, agents/risk-assessor.md,
agents/testability-reviewer.md, agents/requirements-quality-analyst.mdProvides CDSS development patterns for drug interaction checking, dose validation, clinical scoring (NEWS2, qSOFA), and alert classification integrated into EMR workflows.
npx claudepluginhub mikeparcewski/wicked-testing --plugin wicked-testing