docs(architecture): baseline current design (#109)

Capture the brownfield architecture and its relationship to the project vision.\n\n- add the current-state SAD and five retrospective ADRs\n- classify vision vectors against implementation evidence\n- record requirements, risks, gates, and construction handoff

.aiwg/reports/abm-gate-report.md

@@ -0,0 +1,15 @@
+# Architecture Baseline Milestone Gate Report
+
+**Status:** PASS  
+**Date:** 2026-07-20
+
+| Criterion | Status | Evidence |
+| --- | --- | --- |
+| SAD exists and is baselined | PASS | `architecture/software-architecture-doc.md`; current-state, >1,000 words |
+| At least three ADRs | PASS | ADR-001 through ADR-005 |
+| Use-case architectural coverage | PASS | SAD traceability covers UC-001 through UC-005 |
+| Test strategy exists | PASS | `testing/test-strategy.md` |
+| No unmitigated blocking architecture risk | PASS | R-01 has enforced controls; weakening them is explicitly release-blocking |
+| Vision/current-state separation | PASS | ADR-005 and `architecture/vision-alignment.md` |
+
+The architecture is baselined for maintenance and controlled evolution. “Construction ready” means the brownfield team can prioritize alignment work; it does not relabel experimental features as stable.

.aiwg/reports/construction-ready-brief.md

@@ -0,0 +1,44 @@
+# Construction Ready Brief
+
+**Project:** T3MP3ST  
+**Date:** 2026-07-20  
+**Status:** CONSTRUCTION READY — brownfield architecture alignment
+
+## Executive Summary
+
+T3MP3ST is baselined as a local-first TypeScript modular monolith with CLI, library, localhost War Room/API, MCP, provider-neutral reasoning, mission/operator orchestration, deterministic target/arsenal safety controls, evidence/reporting, source analysis, and reproducible evaluation. The architecture documents what is implemented and explicitly refuses to equate `VISION.md` with current capability.
+
+The highest-value construction work is architectural governance around existing code: validate the SAD/ADRs, automate maturity-claim consistency, complete the safety-test inventory for network adapters, and establish workload baselines. No rewrite or microservice migration is implied.
+
+## Artifact Index
+
+| Artifact | Status |
+| --- | --- |
+| Intake form, solution profile, option matrix, risk screening | Baselined |
+| LOM gate | PASS |
+| UC-001–UC-005, user stories, NFR register | Baselined |
+| Current-state SAD | Baselined |
+| ADR-001–ADR-005 | Accepted / retrospective |
+| Vision-to-code alignment matrix | Baselined |
+| Test strategy | Baselined |
+| ABM gate | PASS |
+| Iteration 001, team profile, CI/CD additions | Ready |
+
+## Key Decisions
+
+1. Retain a local-first modular monolith.
+2. Keep reasoning provider-neutral and untrusted.
+3. Enforce scope/approval beneath model reasoning.
+4. Derive public claims from versioned evidence.
+5. Separate implemented current state from research vision.
+
+## Risks to Watch
+
+Scope bypass, secret/evidence leakage, maturity overstatement, nondeterministic false findings, and external-tool supply-chain risk remain the leading concerns. See `intake/risk-screening.md` for controls and residual status.
+
+## First Steps
+
+1. Maintainers review the SAD and mark any ADR amended or superseded.
+2. Execute Iteration 001 architecture-alignment work.
+3. Preserve current CI gates while adding traceability/maturity checks.
+4. Update the alignment matrix whenever a vision capability changes maturity.

.aiwg/reports/lom-gate-report.md

@@ -0,0 +1,14 @@
+# Lifecycle Objective Milestone Gate Report
+
+**Status:** PASS  
+**Date:** 2026-07-20
+
+| Criterion | Status | Evidence |
+| --- | --- | --- |
+| Problem statement defined | PASS | `intake/intake-form.md` |
+| Success metrics defined | PASS | Five measurable criteria in intake |
+| Stakeholders identified | PASS | Five stakeholder groups in intake |
+| Initial risk screening complete | PASS | Seven risks in `intake/risk-screening.md` |
+| Solution approach viable | PASS | Existing operational codebase; no infeasible constraint |
+
+The gate authorizes brownfield elaboration. Unknown adoption, staffing, and SLO facts remain owner inputs but do not prevent documenting current architecture.