Organize the test case IDs in usecase.md into executable sets (smoke / regression / targeted), and specify execution order, dependencies, and blocking criteria。
Inputs and gates (must be followed)
REQUIRED SUB-SKILL: currently executing spec-context to obtain context, and echo back in the conversation FEATURE_DIR=...(allowed (reuse)）。
- Required project-level memory：
project/memory/product.md、project/memory/tech.md、project/memory/glossary.md; if it cannot be read, you must writeCONTEXT GAP。 - Input dependencies：
{FEATURE_DIR}/verification/usecase.mdand{FEATURE_DIR}/verification/test-plan.mdshould already exist; if missing, you must writeCONTEXT GAPand stop (to avoid arranging things out of thin air). - Suites must be locatable: suite entries must point to specific test case IDs (vague descriptions like "cover the login flow" are prohibited).
Output (write to disk)
{FEATURE_DIR}/verification/suites.md
Mandatory template (must be followed)
suites.mdmust be generated according to the template (do not invent your own structure/field names):assets/suites_template.md
Content requirements (minimum structure)
suites.md must contain at least:
- Smoke
- Blocking rules (consistent with the exit criteria of
test-plan.md) - Test case list:
TC-*
- Blocking rules (consistent with the exit criteria of
- Regression
- Test case list:
TC-*(can be grouped)
- Test case list:
- Targeted (targeted regression, if applicable)
- Trigger conditions (impact-analysis / risk list / change description)
- Test case list:
TC-*
- Execution order and dependencies (e.g., smoke → targeted → regression)
DoD self-check (V3-DoD)
- Every entry in the
TC-*ID - The smoke suite has a clear criterion of "failure blocks delivery", and it is consistent with
test-plan.mdConsistent - The targeted suite can be traced back to the impact scope/risk source (if any
requirements/solution.md#impact-analysis）