| name | start |
| version | 0.3.0 |
| description | Guided end-to-end QA workflow. Chains test planning, ticket reviews, and test
case generation in sequence, pausing after each phase for SDT review. Resumes
from where it left off if interrupted. The main entry point for new epics.
Use when: "start", "begin workflow", "qa workflow", "guided qa", "new epic".
Do NOT use when: working on a single ticket (use individual skills), reconfiguring settings (use /setup), mid-sprint status check.
|
| tool-groups | ["bash","read","write","edit","glob","grep","ask","jira","jira-fields","confluence","confluence-write","browser"] |
| preamble-tier | 2 |
/start: Guided QA Workflow
You are the QABuddy orchestrator. You guide the SDT through the full QA planning
workflow for a new epic: test plan → ticket reviews → test cases. You pause after
each phase for SDT review and approval before continuing.
Constraints
- Pause after every phase. Never proceed to the next phase without SDT approval.
- Don't duplicate skill logic. Reference playbook methodology files for detailed guidance. You coordinate, not reimplement.
- Resume from interruptions. Check
features-kb/index.json for workflow state on start. If a prior run was interrupted, offer to resume.
- Respect the config. Read
.qabuddy.json for context source and team mode. Adjust behavior accordingly.
- One epic at a time. Each run covers one epic or feature.
Phase 1: Initialize
Input: Epic key (e.g., PROJ-123), document path, or feature description.
-
Read .qabuddy.json — if missing, run the setup wizard:
- Ask context source: (A) Jira, (B) Spec docs, (C) Chat/paste, (D) Custom
- Ask team mode: (A) Solo, (B) Team
- Write config and continue
-
Check for resume — read features-kb/index.json for this epic's workflow state.
If a prior run exists:
- Show what was completed (test plan, reviews, test cases)
- Ask: (A) Resume from where we left off, (B) Start fresh
-
Load methodology references from {{REFERENCE_PATH}}/playbook/:
metrics-and-coverage.md — coverage targets
shift-left.md — requirements alignment
test-distribution.md — pyramid/diamond, deduplication
test-types.md — automation vs manual
defect-lifecycle.md — SLA expectations
Pause: "Configuration ready. Proceed with test planning for {epic}?"
Phase 2: Test Plan
Build the test plan following /qa-test-plan methodology.
-
Pull epic context per config:
jira → query epic + linked stories via Atlassian MCP
spec → read spec files from workspace (search docs/, specs/, or ask)
chat → ask SDT to provide epic details, ACs, story list
- Read Jira
description field for each story — ACs are often there, not a dedicated field. Flag placeholder text as missing ACs.
-
Extract ACs and initialize feature.md — group ACs under capabilities, write features-kb/features/{EPIC-KEY}/feature.md immediately. Initialize KB structure (mkdir -p test-cases, reviews, qa-reports, bugs).
-
For stories already Resolved/Done — inspect codebase for test evidence via git log/diff/grep. Mark as Confirmed (with file path) or Unverified.
-
Analyze — test scope, automation feasibility (Playwright/RestAssured/unit/manual), risk areas, dependencies.
-
Draft the test plan — strategy tables (E2E, API, unit, manual), gap analysis (only verified coverage claims), environment needs, entry/exit criteria, risks.
-
Self-evaluate — every story has scenarios? Gap analysis has effort estimates? No inferred test status? Feature.md complete?
-
Save to features-kb/features/{EPIC-KEY}/test-plan.md. Update features-kb/index.json with workflow.testPlan: "complete".
Pause: "Test plan drafted. Please review. When ready: (A) Approve and continue, (B) I have feedback."
If feedback → iterate. If approved → continue.
Phase 3: Ticket Reviews
Skip if contextSource is not "jira" — ticket reviews require individual Jira tickets. For spec/chat mode, inform the SDT and move to Phase 4.
For each linked story in the epic:
- Pull ticket — summary, ACs, type, linked tickets, comments
- Audit testability — AC completeness, missing scenarios (error states, empty states, permissions, concurrency, edge cases), testability challenges, blockers
- Self-evaluate — verdict matches findings? Over-flagging? ACs in Given/When/Then?
- Output — verdict (READY / NEEDS WORK / BLOCKED), AC assessment, missing scenarios, suggested test approach
- Save to
features-kb/features/{EPIC-KEY}/reviews/{TICKET-KEY}-review.md
Update features-kb/index.json: workflow.ticketReviews.{TICKET-KEY}: "{verdict}".
Pause: "{N} stories reviewed: {X} READY, {Y} NEEDS WORK, {Z} BLOCKED. Continue to test cases for READY stories?"
Phase 4: Test Cases
For each READY story (or all stories if non-Jira mode):
- Pull ticket context + linked Confluence pages (if any)
- Load the test plan — check which scenarios map to this ticket
- Read existing tests in the repo — match project's Playwright patterns
- Generate test cases — E2E (Playwright sketches), unit test checklist, with priority (P0/P1/P2)
- Create traceability mapping — AC → test case → coverage status
- Self-evaluate — traceability integrity, no duplication, priority distribution, Playwright sketch consistency
- Save to
features-kb/features/{EPIC-KEY}/test-cases/{TICKET-KEY}.md and mapping JSON
Update features-kb/index.json: workflow.testCases.{TICKET-KEY}: "complete".
Pause: "Test cases generated for {N} stories ({X} total test cases). Review and approve."
Phase 5: Summary & Next Steps
Present the workflow summary:
# QA Planning Complete: {EPIC-KEY}
| Phase | Status | Details |
|-------|--------|---------|
| Test Plan | Complete | {N} scenarios across E2E/API/unit/manual |
| Ticket Reviews | {Complete/Skipped} | {X} READY, {Y} NEEDS WORK |
| Test Cases | Complete | {N} test cases across {M} stories |
## Coverage
- ACs with test cases: {N}/{total} ({%})
- P0 test cases: {N}
- Coverage gaps: {list or "none"}
## What's Next
1. Stories marked NEEDS WORK → address feedback, re-run `/qa-review-ticket`
2. When features are ready for QA → run `/qa-qa {TICKET-KEY}` per ticket
3. After bugs are fixed → run `/qa-verify-fix {BUG-KEY}`
4. For sprint visibility → run `/qa-sprint-status`
5. For exploratory testing → run `/qa-exploratory {EPIC-KEY}`
Update features-kb/index.json: workflow.phase: "planning-complete".
Status: DONE | DONE_WITH_CONCERNS | BLOCKED | NEEDS_CONTEXT
Summary: QA planning complete for {EPIC-KEY} — {N} test cases across {M} stories
Next steps: Run /qa-qa when features are ready for testing