Exploratory Testing Workflow

You define exploratory testing intent before the suite moves into detailed planning or automation.

Skills

Activity	Skill
Define exploratory charter, risk hypotheses, investigation order, and exploration gaps	exploratory-test-writer
Gather grounded product evidence	capture-feature-evidence
Turn confirmed risks into shared coverage planning and final P0/P1/P2 prioritization	plan-test-coverage
Turn shared plans into detailed specs	generate-test-cases

Workflow Sequence

The default progression is:

write exploratory charter → run explore app when needed to close evidence gaps → finalize charter → optionally plan coverage → optionally generate specs

Orientation

• Load exploratory-test-writer as the intent-layer entry point.
• Use the first charter pass to decide whether grounded product evidence is missing, stale, or too thin to support confident risk framing.
• If the charter depends on grounded product behavior and e2e-plan/exploration-report.md is missing or stale, run capture-feature-evidence, then return to exploratory-test-writer to finalize the charter.
• Use the completed charter to identify what still needs deeper probing and what should be handed to shared planning next.

Handoff Rules

• exploratory-test-writer owns e2e-plan/exploratory-charter.md, including mission, risk hypotheses, exploratory ordering, and evidence gaps.
• plan-test-coverage may consume that charter as optional context, but it owns the final P0/P1/P2 coverage priority and it does not let the charter replace the required grounded evidence in e2e-plan/exploration-report.md.
• capture-feature-evidence owns e2e-plan/exploration-report.md.
• plan-test-coverage owns e2e-plan/coverage-plan.md.
• generate-test-cases owns test-cases/<feature>.md.

If the user explicitly asks for shared coverage planning or test-case specs, load the shared test-analysis skills after the charter is complete. Do not encode that sequencing inside the plugin skill itself.

If runnable automation is requested after the exploratory charter is finalized, stop with an explicit recommendation to move into the appropriate execution workflow:

• playwright-automation for Playwright
• robot-automation for Robot Framework

This workflow does not load execution-layer plugins itself, so it must not claim to implement or run framework automation in the same workflow session.

Guardrails

• Do not present inferences as observed evidence.
• Do not write framework-specific automation from this workflow.
• Stop and recommend deeper exploration if the charter depends on evidence that has not been gathered yet.