testgen-flow

<testgen_flow>

<description_and_purpose>

Systematic requirements analysis from Issue Tracker tickets and Wiki documentation to structured requirements and test scenarios. Extracts data, identifies gaps, clarifies unknowns via HITL, generates requirements document, and produces test cases with export to a Test Management System (Phase 6, user-triggered — the user may choose not to trigger it, but it is a fully-specified phase, not informally skippable). Designed for BA/QA engineers and requirements engineers.

Prerequisite: Rosetta Prep Steps.

Terminology. External systems are named by role throughout this workflow and its phases: Issue Tracker, Wiki, and Test Management System (TMS). Jira, Confluence, and TestRail are canonical examples only — adapt identifiers, URLs, requests, calls, and query syntax to the systems resolved for the current project (from repository-root gain.json, explicit user input, recognizable URLs/handles, and available integrations).

</description_and_purpose>

<workflow_phases>

• All Rosetta prep steps MUST be FULLY completed
• USE SKILL load-project-context, orchestration, hitl
• ONE PHASE AT A TIME: APPLY PHASE the phase file, execute, update state, move to next. JIT: phase-by-phase: do not load/act IN ADVANCE.
• DO NOT SKIP PHASES: Each builds on the previous. Skip gates: only with explicit user instruction, or when testgen-state.md marks the phase complete and its expected output file exists under plans/testgen-{TICKET-KEY}/; otherwise resume from the earliest incomplete phase. The explicit user instruction skip NEVER applies to the Phase 3 / Phase 6 HITL gates — those are rule 2 of <orchestration_and_escalation> and are never overridden.
• Phase-file load failure: if APPLY PHASE for a phase file returns nothing, retry once, then HALT and report — do not improvise the phase.
• Transition precedence: <orchestration_and_escalation> priority hierarchy.
• STATE TRACKING: Update plans/testgen-{TICKET-KEY}/testgen-state.md after each phase.
• SELF-CHECK BETWEEN PHASES: Before advancing, verify the state row was updated, the expected output file exists and is non-empty, the phase's ## Metrics count is populated (a thin 0/1 → re-check the artifact), and any HITL approval (Phase 3, 6) is recorded.
• When a phase delegates work to subagents, dispatch per USE SKILL orchestration.
• MUST ALWAYS use todo tasks ledger, ASAP. Prioritize ACCURACY over SPEED. Proactively use available tools/MCPs, if question can be answered from code or MCP call - do that first.
• MUST create output directory plans/testgen-{TICKET-KEY}/ at start.
• Trigger prompt example: Analyze requirements for PROJ-123 (also: bare key PROJ-123, full ticket URL). Ticket-only and ticket+Wiki input formats are enumerated in testgen-flow-project-config-loading.md step 0.1.
• Per-phase failure cases — owned by phase files:
  • Ticket not found → testgen-flow-data-collection.md + the data-collection skill's Issue Tracker failure handling.
  • No Wiki results → testgen-flow-data-collection.md + the data-collection skill's Wiki failure handling.
  • User declines / does not answer questions → testgen-flow-question-generation.md <failure_handling> "User explicitly declines to answer".
  • Incomplete / missing requirements inputs → testgen-flow-requirements-document-generation.md <failure_handling> "Missing or empty inputs".
  • Documentation search query example + ranking rule → the data-collection skill's documentation-binding search/ranking behavior.
  • Initial-prompt format examples → testgen-flow-project-config-loading.md.
  • Phase 6 80%-export-success threshold → testgen-flow-test-case-export.md (Threshold (80%) met field + PARTIAL — N/M exported state).
  • Phase 5 test-case-count guidance → testgen-flow-test-case-generation.md <validation_checklist>.
• Model tiers (phase subagent_required_model): tier: complex = heavy reasoning / multi-source synthesis / requirements engineering (Opus-class / GPT high-tier); tier: workhorse = structured execution / extraction / generation + export (Sonnet-class / GPT mid-tier).