Estimate Assemble

Output: Write estimation-infra.json

Assemble the full artifact conforming to references/vendored/estimate/estimation-infra.schema.json (that schema is the field contract — do not re-enumerate it here). Each section is the corresponding output the cost-engine fragment computed: pricing_source + current_costs (Part 1), projected_costs (Parts 2/2B), cost_comparison (Part 3), migration_cost_considerations (Part 4), roi_analysis (Part 5), optimization_opportunities (Part 6), complexity_tier + complexity_inputs (Part 7), and recommendation (Part 8).

The assembler additionally DERIVES the financial_summary roll-up (not produced by any single cost-engine Part; the schema leaves its shape open):

Sign convention: savings = Heroku minus AWS (positive = you save by migrating). This is deliberately the OPPOSITE sign of roi_analysis.recurring_savings.monthly_difference_* (difference = AWS minus Heroku) — same fact, savings-vs-difference framing. When presenting either, always label the direction in words; never print a bare signed value.

Also attach optional workshop metadata when present (does not affect Property-16):

Write to $MIGRATION_DIR/estimation-infra.json.

Completion Handoff Gate (Fail Closed)

The completion checks are declared in this phase’s _postconditions frontmatter and enforced per INTERPRETER.md § Gate protocol: re-read estimation-infra.json from disk, run the mechanical checks (_check_file_exists / _validate_json) and the _assert judgment checks (recommendation shape, the Property-16 total-invariant, every-service-priced, complexity tier), then emit GATE_FAIL (do NOT patch artifacts; STOP) or HANDOFF_OK | phase=estimate | artifacts=estimation-infra.json.

One check needs this fragment’s context: estimation-infra.json must also pass references/vendored/estimate/estimation-infra.schema.json validation (the schema shape) — verify that as part of the _validate_json postcondition.

Inner workshop reprice — skip this gate’s state transition

When invoked from workshop-refresh.md (inner reprice): write estimation-infra.json, optionally soft-check Property-16, present a brief summary, then return to the workshop loop. Do not emit HANDOFF_OK, do not update .phase-status.json, do not offer the what-if workshop below.

Present Summary

After writing estimation-infra.json, present a concise summary to the user:

1. Pricing source and accuracy — State cache age and accuracy range
2. Heroku baseline vs AWS projected (balanced tier) — one-line comparison (if a baseline was determined, labeled with its source; include the derived-baseline caveat when the source is not billing data)
3. Three-tier table: Premium, Balanced, Optimized with monthly totals
   • Premium: Highest resilience / highest monthly estimate
   • Balanced: Default scenario; compare Heroku to this first
   • Optimized: Lower estimate; reservations / Spot trade-offs assumed
   • One-line note: Three figures are pricing scenarios for the same architecture (not three Terraform stacks). Generated Terraform aligns with Balanced.
4. Per-service cost breakdown (balanced tier, 1 line per service)
5. Migration complexity: tier + timeline range
6. Monthly and annual savings (or increase) vs Heroku per tier (if a baseline was determined)
7. Cost optimization: if optimization_opportunities is non-empty, list the top 2-3 with savings potential. If it is empty (per the vendored eligibility matrix, nothing in this design is RI/SP-eligible), say so in one line instead of skipping the line entirely: “No 1-year/3-year commitment product applies to this architecture.” Either way, when any commitment product is mentioned, add: “Note: Activate credits don’t cover RI/Savings Plan upfront costs.”
8. Recommendation: lead with outcome_label (or path_label when outcome is absent). List conditions[] when conditional_go. Close with 1–3 would_flip_if[] bullets when present. Keep path_label as the execution-shape line under the verdict.

Keep under 25 lines. The user can ask for details or re-read estimation-infra.json.

Phase status after outer Estimate (deferred Generate advance)

After outer-run HANDOFF_OK (not an inner workshop reprice):

1. Mark phases.estimate → "completed".
2. Ensure phases.workshop exists (seed "pending" if the key is missing).
3. Do not set current_phase to "generate" yet — leave current_phase at "estimate" until the workshop sidebar is resolved (entered then exited, or declined) and the Decision gate below has been answered. This matches sidebar semantics: workshop never owns current_phase, and mid-workshop fixtures correctly stay on estimate.
4. Offer the what-if workshop below. On exit or decline, present the Decision gate (below) — do not fall through to Generate directly.

Post-Estimate: What-If Workshop Offer

After outer-run HANDOFF_OK, the summary above, and the deferred phase-status update — offer:

Data-justified scenario hint (add one line when applicable): if a material assumption was defaulted or tier-derived rather than confirmed — most commonly database_ha — append: “Suggestion: we assumed [assumption]; comparing a [alternative] scenario would bound it before you commit.” Suggest at most one.

• A → Load references/phases/workshop/workshop.md (sidebar) and follow it (baseline capture if scenarios/ missing, then the sheet). Keep current_phase: estimate; set phases.workshop → "in_progress". On workshop exit, return to the Decision gate below (do not advance to Generate directly) — the workshop’s active scenario carries into it.
• B → Mark phases.workshop → "completed" (resolved/declined — no scenarios/ required). Proceed to the Decision gate below.

On first workshop entry after this Estimate, workshop-refresh.md baseline capture snapshots the current artifacts as scenario-001 before any edits.

Post-Estimate: Decision Gate

The decision is the product; execution artifacts are opt-in. The verdict (recommendation.outcome / path) already exists in estimation-infra.json — present it and let the user choose what happens next. Never advance to Generate without an explicit choice of option C (or an explicit later request for Terraform/scripts).

Reached only after the what-if workshop offer above has been resolved (entered-and-exited, or declined) — never presented while phases.workshop is "pending" or "in_progress".

Present (values from estimation-infra.json; one line each):

Omit option B if the workshop sidebar is already "completed" from the offer above (do not re-offer the same choice twice in one turn) — present only [A] Done for now and [C] Generate Terraform and migration scripts in that case.

Choice handling:

• A → Then:
  1. Render the decision pack: load references/shared/report-decision-core.md and render it in decision mode — write $MIGRATION_DIR/decision-report.html and $MIGRATION_DIR/DECISION.md per that file’s decision-mode rules (no appendices, no Terraform, CTA footer). Validate with python3 "<SKILL_BASE>/scripts/validate-heroku-migration-report.py" "$MIGRATION_DIR/decision-report.html" --mode decision --migration-dir "$MIGRATION_DIR" (absolute paths — cwd must not be load-bearing; --migration-dir is required so the decision-mode pre-execution check — .phase-status.json‘s phases.generate must be "pending" or absent for THIS cycle, not raw terraform/ / generation-*.json absence; see report-decision-core.md — actually runs) and fix failures before presenting.
  2. Set run_mode: "decide" and current_phase: "complete" in .phase-status.json (phases.generate stays "pending" — this combination means “decision complete, execution available on request”; see references/vendored/state/phase-status.schema.json).
  3. Continue with the Feedback sidebar per SKILL.md. Close with: “Your decision report is saved at decision-report.html (plus a Slack-friendly DECISION.md). If you decide to migrate, say ‘generate the Terraform and migration scripts’ — everything is saved and I’ll pick up from here.”
• B → Load references/phases/workshop/workshop.md. Keep current_phase: estimate; set phases.workshop → "in_progress". On workshop exit, return to this gate (options A and C; the workshop’s active scenario carries into either) — do not advance to Generate directly.
• C → Set run_mode: "decide_and_execute" and current_phase → "generate". Continue with the Feedback/Generate sidebars in SKILL.md.

Decide-complete resume

If a warm start finds current_phase == "complete" AND run_mode == "decide" AND phases.generate == "pending": this is the decide-complete terminal state, not an incomplete run. Do not re-run Estimate. Offer:

• A → Set run_mode: "decide_and_execute" and current_phase → "generate" before loading generate.md. Continue to Generate.
• B → Leave state unchanged; end the turn.