XK population QA cycle 1: review bundle

Task metadata

Latest summary

PASS: reviewed XK cycle-1 bundle at /home/synthestat/output/runs/XK/xk_population_review_cycle1_3bf65dd8_seed420987; it is fit for internal research review of the declared seeded Kosovo slice, with no critical failures. Full report: /home/synthestat/workspace/manager_handoffs/reviewer/2026-05-19_164813Z_XK_population_review_cycle1.md

Body

Review the Kosovo (XK) population synthesis cycle-1 bundle produced by parent synth-modeler task.

Project root: /home/synthestat
Country: XK
Release mode: internal research review
Contract: /home/synthestat/docs/contracts/population_review_bundle.md
Parent handoff to inspect: modeler task t_65697679

Required context to read before review:
- /home/synthestat/docs/SOUL.md
- /home/synthestat/docs/contracts/population_review_bundle.md
- /home/synthestat/docs/dashboard/countries/XK.md
- Parent task handoff and bundle path from kanban_show parent context

Review requirements:
1. Validate that the review bundle is contract-complete. Missing required files must be either present or have explicit unavailable.json reasons where the contract allows.
2. Check non-negotiables:
   - HARD constraints exact; FIRM/SOFT residuals within declared tolerance or explicitly relaxed.
   - No hidden-population overlay silently rewrites de jure constraints.
   - Every modelled estimate has uncertainty bounds or explicit unavailable/degraded status.
   - Geography quality tiers, degraded zones, relaxed constraints, missing sources, and failed layers are explicit.
   - Provenance includes source IDs/URLs, retrieval timestamps where available, geography levels, reference periods, and quality flags.
3. Treat Kosovo as sparse-evidence / lower-tier: relaxed constraints and wider intervals are acceptable only when visible in metadata and QA, not when hidden as precision.
4. Evaluate whether the current output is fit for internal research review, not external release.

Output contract:
Return exactly one routing verdict in metadata and summary:
- PASS
- NEEDS_MODEL_FIX
- NEEDS_MORE_SOURCES
- BLOCKED_INVALID_OUTPUT
- EVIDENCE_EXHAUSTED_HUMAN_REVIEW
- MODEL_IMPROVEMENT_EXHAUSTED_HUMAN_REVIEW

Also include:
- bundle_path reviewed
- decisive findings
- required next action if not PASS
- whether this failure is materially similar to any prior XK cycle (if prior cycles exist)

Do not fix/model/download/research yourself. Review only.

Runs

Events

Comments

Reviewer handoff written to `/home/synthestat/workspace/manager_handoffs/reviewer/2026-05-19_164813Z_XK_population_review_cycle1.md`.

Verdict: PASS

Decisive findings:
- Bundle reviewed: `/home/synthestat/output/runs/XK/xk_population_review_cycle1_3bf65dd8_seed420987`.
- Contract-complete with required parquet/json/md artifacts or explicit unavailable markers for hidden-population and work/school assignment layers.
- HARD constraints pass exactly; FIRM residuals pass with max_abs_relative_error=0.0; no relaxed constraints.
- Internal link checks passed for person-household, household-dwelling, and dwelling-building assignment references.
- Weak layers and sparse-evidence limitations are explicit: seeded 2-zone slice only, no hidden-population overlay, no work/school/facility assignment, modelled/transfer distributions visible with uncertainty.
- No prior XK reviewer cycle/failure found in manager handoffs; this is not materially similar to a prior XK review failure.