Revert/feedthrough kirchhoff bandaid

.gitignore

@@ -34,6 +34,3 @@ dmypy.json
 coverage_output.log
 **/.DS_Store
 .local_coverage_data
-
-# Local credentialed runner for validate_lug_derivation
-scripts/validate_lug_derivation/run_local.sh

CHANGELOG.md

@@ -4,18 +4,6 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [2.6.3] - 04/2026
-
-### Fixed
-
-- **Feedthrough values now derived via Kirchhoff instead of read from `downstream-lugs`** — `SpanPanelSnapshot.feedthrough_power_w`, `feedthrough_energy_consumed_wh`, and `feedthrough_energy_produced_wh` are computed as `main − Σ(branches)` inside
-  `HomieDeviceConsumer._build_snapshot` (and mirrored in the dirty-circuit rebuild path) rather than sourced from the native `energy.ebus.device.lugs.downstream` `active-power` / `imported-energy` / `exported-energy` properties. Kirchhoff at the main bus
-  produces a physically-consistent result: `P_main = P_feedthrough + Σ(branches, load-perspective)` for instantaneous power, and the same identity applied to _net_ energy (`main.consumed − main.produced − Σ(branch.net)`) split into non-negative consumed /
-  produced counters. Net-based energy handling is required because a per-direction subtraction would emit negative cumulative counters whenever circuits flow bidirectionally (the classic case is PV self-consumption: `Σ(branch.consumed)` can exceed
-  `main.consumed` even when the net balance is valid). The synthesized PV virtual circuit participates with the correct load-perspective sign, and unmapped tab entries are zero-power, so both contribute safely to the sum. No public interface change — field
-  names and types are unchanged; only the source of the values shifts. `downstream_l1_current_a` / `downstream_l2_current_a` continue to be read directly from the downstream-lugs node — those per-phase readings are orthogonal to the defect. The underlying
-  firmware defect is tracked upstream at [spanio/SPAN-API-Client-Docs#13](https://github.com/spanio/SPAN-API-Client-Docs/issues/13).
-
 ## [2.6.2] - 04/2026
 
 ### Changed

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "span-panel-api"
-version = "2.6.3"
+version = "2.6.2"
 description = "A client library for SPAN Panel API"
 authors = [
     {name = "SpanPanel"}