reconstruction-tracker

name	reconstruction-tracker
description	This skill should be used when the user asks to "track reconstruction progress", "compare pre-conflict vs current cadastre", "damage assessment for this area", "find destroyed buildings", "post-conflict reconstruction monitoring", "map damaged parcels", "show reconstruction status", or invokes `/civic-gis:reconstruction-tracker`. Compares a pre-event cadastre baseline (parcels + buildings before conflict / disaster) with a post-event survey or remote-sensing layer, classifies each parcel as intact / damaged / destroyed / rebuilt, and produces a per-parcel status report plus aggregate statistics. Composes on top of gis-to-db:analyze-diff.
argument-hint	--baseline path/to/pre-event.geojson --current path/to/post-event.geojson [--damage-layer path/to/damage.geojson] [--key parcel_id]
allowed-tools	Bash, Read, Write, Glob

Reconstruction Tracker

v0.1 status — SKILL.md only. Composes on top of gis-to-db:analyze-diff for the parcel-by-parcel comparison primitive. Full Python orchestrator lands in v0.2.

Compare pre-event cadastre with current state and classify each parcel's status. Used for post-conflict reconstruction planning, disaster recovery, and tracking rebuild progress over time.

When to Use

• /civic-gis:reconstruction-tracker --baseline cadastre-2010.geojson --current cadastre-2025.geojson
• "Compare the pre-war cadastre with the current survey of Homs district A."
• "Find destroyed buildings in this area."
• "Map reconstruction progress over the last year."

Status Classification

For each parcel matched between baseline and current:

Status	Definition
intact	Footprint unchanged, building still standing per damage layer (if supplied)
damaged	Footprint mostly intact but flagged in damage layer
destroyed	Footprint missing or <30% of baseline area remaining
rebuilt	Footprint differs from baseline but a new building exists
vacant	Parcel exists in baseline but no building in current
new_construction	Parcel/building in current that was not in baseline
unknown	Match failed; needs manual review

How It Works

1. Validator agent runs first (civic-preflight-validator, mode=reconstruction-tracker).
2. Invoke gis-to-db:analyze-diff on baseline + current to get added/removed/changed/unchanged feature lists.
3. Optionally overlay a damage-classification layer (from drone survey, satellite analysis, or field assessment). Damage layer features tagged with severity (light/moderate/severe/destroyed) override default classification.
4. Per-parcel: assign one of the 7 status values based on the diff result + damage overlay.
5. Aggregate: counts per status, total area per status, hotspot map (clusters of destroyed parcels).
6. Render markdown report + write per-parcel GeoJSON output for visualization.

Inputs

• Baseline GeoJSON — pre-event cadastre. Each parcel must have a stable parcel_id.
• Current GeoJSON — recent survey or remote-sensing-derived footprints. Same parcel_id schema when possible.
• Damage layer (optional) — GeoJSON polygons with damage_severity property.

Output

Markdown report with:

• Summary table: parcel count per status, total area per status, %.
• Top-10 damaged neighborhoods (clusters of destroyed parcels).
• Recommendations: priority zones for restitution/reconstruction.

Plus a per-parcel labeled GeoJSON (<output_dir>/reconstruction-status.geojson) for downstream mapping.

v0.2 Plans

• Pre-built Python orchestrator.
• Integration with UN-Habitat damage classification standards.
• Time-series mode (compare 3+ snapshots to track rebuild velocity).
• Property restitution workflow tying status → displaced-person registry.