diff --git a/evaluations/fold-04-holdout-5/AGENT_ACTION_LOG.md b/evaluations/fold-04-holdout-5/AGENT_ACTION_LOG.md
new file mode 100644
index 0000000..b4aeaf3
--- /dev/null
+++ b/evaluations/fold-04-holdout-5/AGENT_ACTION_LOG.md
@@ -0,0 +1,24 @@
+- 2026-06-30T15:54:33-07:00 | action=command | subject=`mkdir -p /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/agent_outputs` | reason=initialize benchmark output directory and action log
+- 2026-06-30T15:54:42-07:00 | action=read | subject=`/scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/{AGENT_INSTRUCTIONS.md,MANIFEST.json,skills/essdive_sm_curator/SKILL.md,skills/essdive_sm_harmonizer/SKILL.md}` | reason=load fold-local task instructions and allowed skills
+- 2026-06-30T15:54:51-07:00 | action=read | subject=`/scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/skills/essdive_sm_curator/SKILL.md lines 261-520 and skills/essdive_sm_harmonizer/SKILL.md lines 261-620` | reason=complete fold-local skill instructions before acting
+- 2026-06-30T15:54:58-07:00 | action=command | subject=`wc -l /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/skills/essdive_sm_curator/SKILL.md /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/skills/essdive_sm_harmonizer/SKILL.md` | reason=verify whether skill reads reached end of files
+- 2026-06-30T15:55:06-07:00 | action=read | subject=`/scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/skills/essdive_sm_curator/SKILL.md lines 521-556` | reason=finish reading fold-local curator skill
+- 2026-06-30T15:55:13-07:00 | action=directory_listing | subject=`find /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5 -maxdepth 5 -type f` | reason=identify fold-local allowed exemplars and metadata files
+- 2026-06-30T15:55:23-07:00 | action=read | subject=`/scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/data/processed/ess-dive_wfsfa_soil_datasets/sm_data_harmonization_mapping.json and selected fold-local expert code` | reason=inspect allowed schema examples and code patterns for harmonizer output
+- 2026-06-30T15:55:35-07:00 | action=directory_listing | subject=`find /h/jmc/ess-dive_wfsfa_soil_datasets/ess-dive-8ac2940c708a515-20230504T210140482233 -maxdepth 3 -type f` | reason=inspect allowed held-out raw package files
+- 2026-06-30T15:55:42-07:00 | action=directory_listing | subject=`find /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/data -maxdepth 6 -type f` | reason=locate fold-local cached metadata and allowed exemplar files
+- 2026-06-30T15:56:34-07:00 | action=read | subject=`python inspect CSV headers and previews in /h/jmc/ess-dive_wfsfa_soil_datasets/ess-dive-8ac2940c708a515-20230504T210140482233` | reason=classify raw held-out payload, sensor metadata, and units
+- 2026-06-30T15:56:55-07:00 | action=read | subject=`python summarize /h/jmc/ess-dive_wfsfa_soil_datasets/ess-dive-8ac2940c708a515-20230504T210140482233 CSVs` | reason=determine held-out payload row counts, unique sites, sensors, depths, and interval
+- 2026-06-30T15:57:10-07:00 | action=search | subject=`rg -n "Microclimate|SM_SWC|SFA_Location|Sensor.SN|Treatment|Subplot|SWC" /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/data` | reason=find allowed exemplar patterns for this raw file structure
+- 2026-06-30T15:57:19-07:00 | action=read | subject=`/scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/data/gold/expert_code/harmonize_sm/dataset_23.py and mapping lines 1500-1635` | reason=inspect allowed analogous treatment-filtered SWC harmonization pattern
+- 2026-06-30T15:57:32-07:00 | action=directory_listing | subject=`find /h/jmc/ess-dive_wfsfa_soil_datasets/ess-dive_b924ba040ecefee_20250707T154402200 -maxdepth 4 -type f` | reason=locate allowed raw reference location registry cited by fold-local mapping index 0
+- 2026-06-30T15:57:46-07:00 | action=read | subject=`python inspect /h/jmc/ess-dive_wfsfa_soil_datasets/ess-dive_b924ba040ecefee_20250707T154402200/data/East_Taylor_Watershed_Community_Observatory_Sites___Point_Locations__Surface_v3_2_20250327.csv and match held-out SFA_Location values` | reason=resolve held-out site coordinates using allowed raw location registry
+- 2026-06-30T15:58:02-07:00 | action=read | subject=`python search matched prefixes in /h/jmc/ess-dive_wfsfa_soil_datasets/ess-dive_b924ba040ecefee_20250707T154402200/data/East_Taylor_Watershed_Community_Observatory_Sites___Point_Locations__Surface_v3_2_20250327.csv for missing held-out locations` | reason=resolve or document unmatched location IDs
+- 2026-06-30T15:58:27-07:00 | action=read | subject=`python dry-run held-out control filter and coordinate coverage for /h/jmc/ess-dive_wfsfa_soil_datasets/ess-dive-8ac2940c708a515-20230504T210140482233` | reason=validate harmonization decisions before writing outputs
+- 2026-06-30T15:58:45-07:00 | action=command | subject=`curl -L https://api.ess-dive.lbl.gov/packages/ess-dive-8ac2940c708a515-20230504T210140482233` | reason=retrieve public ESS-DIVE metadata DOI because no fold-local cache is present
+- 2026-06-30T15:58:54-07:00 | action=command | subject=`curl -L https://api.ess-dive.lbl.gov/packages/ess-dive-8ac2940c708a515-20230504T210140482233` | reason=retry public ESS-DIVE metadata DOI lookup after sandbox DNS failure
+- 2026-06-30T16:01:05-07:00 | action=write | subject=`python generate curator_bundle.json mapping.json mapping_notes.md harmonize_heldout.py in /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/agent_outputs` | reason=create required benchmark output artifacts from allowed raw files and fold-local exemplars
+- 2026-06-30T16:01:26-07:00 | action=command | subject=`python /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/agent_outputs/harmonize_heldout.py` | reason=run generated harmonizer to create heldout_harmonized.csv
+- 2026-06-30T16:01:48-07:00 | action=read | subject=`python validate files in /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/agent_outputs` | reason=verify required output files, schema, row count, and basic harmonized values
+- 2026-06-30T16:02:05-07:00 | action=command | subject=`python -m py_compile /scratch/jmc/data-harmonization-eval/.runs/fold-04-holdout-5/agent_outputs/harmonize_heldout.py` | reason=syntax-check generated harmonization script
+- 2026-06-30T16:02:54-07:00 | action=command | subject=`python action-log forbidden-path scan` | reason=verify trace contains no forbidden outside-fold access patterns
diff --git a/evaluations/fold-04-holdout-5/AGENT_INSTRUCTIONS.md b/evaluations/fold-04-holdout-5/AGENT_INSTRUCTIONS.md
new file mode 100644
index 0000000..85c17b4
--- /dev/null
+++ b/evaluations/fold-04-holdout-5/AGENT_INSTRUCTIONS.md
@@ -0,0 +1,10 @@
+# Run environment — leave-one-cluster-out
+
+Harmonize the held-out dataset(s) below using ONLY:
+- the skills in `skills/`,
+- the exemplars in `data/processed/.../sm_data_harmonization_mapping.json` and the code patterns in `data/gold/expert_code/harmonize_sm/` (both have the held-out cluster removed),
+- the shared raw inputs under `~/ess-dive_wfsfa_soil_datasets/` and the cached metadata under `data/external/ess-dive_meta/`.
+
+Do NOT look up the held-out dataset's existing harmonized output, expert code, or mapping entry from any other location. The held-out datasets are:
+
+- `ess-dive-8ac2940c708a515-20230504T210140482233`
diff --git a/evaluations/fold-04-holdout-5/EVALUATION.md b/evaluations/fold-04-holdout-5/EVALUATION.md
new file mode 100644
index 0000000..39f7f1a
--- /dev/null
+++ b/evaluations/fold-04-holdout-5/EVALUATION.md
@@ -0,0 +1,65 @@
+# Fold 04 Holdout 5 Agent Evaluation
+
+## Scope
+
+This directory records one sandboxed sub-agent attempt on fold 04, where dataset index 5 was held out.
+
+- Dataset index: `5`
+- Sandbox: `.runs/fold-04-holdout-5`
+- Tracked artifact copy: `evaluations/fold-04-holdout-5`
+- Successful agent trace: `AGENT_ACTION_LOG.md`
+
+## Anti-Leakage Audit
+
+The successful trace was scanned for root gold data, root processed data, other fold sandboxes, `dataset_05.py`, and explicit `violation` markers. Result: `0` matches.
+
+## Output Equivalence Against Gold
+
+| Metric | Agent | Gold |
+| --- | ---: | ---: |
+| Rows | 515,825 | 1,089,621 |
+| Columns | 9 | 9 |
+| Schema exact match | true | n/a |
+| Unique keyed rows recovered | 515,795 | 976,805 |
+| Missing gold keyed rows | 461,010 | n/a |
+| Extra agent keyed rows | 0 | n/a |
+| Row precision | 1.000 | n/a |
+| Gold recall | 0.528 | n/a |
+
+Shared-row value matches after key merge:
+
+| Column | Matched rows |
+| --- | ---: |
+| `interval_min` | 59007/516005 |
+| `volumetric_water_content_m3_m3` | 515933/516005 |
+| `gravimetric_water_content_gH2O_gs` | 516005/516005 |
+| `water_potential_kPa` | 516005/516005 |
+| `is_timeseries` | 516005/516005 |
+
+Interpretation: this run did not match the gold row set. The agent retained control-treatment SWC rows only, producing 515,825 rows versus 1,089,621 gold rows. Because `depth_m` is missing and timestamps/sites repeat, the key merge has duplicate expansions; row-set metrics above use NaN-safe unique keys.
+
+## JSON Mapping Match
+
+- Top-level exact matches: `5/7`
+- Gold mapping categories present: `8/8`
+- Extra agent categories: `gravimetric_water_content`
+- Missing gold categories: `none`
+- Field-level exact matches: `24/40`
+
+Field exact matches by name:
+
+| Field | Exact matches |
+| --- | ---: |
+| `destination_variable` | 8/8 |
+| `source_pattern` | 4/8 |
+| `source_files` | 4/8 |
+| `transformation` | 0/8 |
+| `unit_conversion` | 8/8 |
+
+## Python Match
+
+The generated script produced the required target schema and parsed successfully, but its curation choice diverged from gold by filtering to `Treatment == "ctl"` rows only. Gold includes a much larger row set for this dataset.
+
+## Overall Assessment
+
+The fold 04 run passed the local no-gold trace audit and generated executable artifacts, but output equivalence was poor because the row-inclusion decision diverged from gold.
diff --git a/evaluations/fold-04-holdout-5/MANIFEST.json b/evaluations/fold-04-holdout-5/MANIFEST.json
new file mode 100644
index 0000000..871b6b8
--- /dev/null
+++ b/evaluations/fold-04-holdout-5/MANIFEST.json
@@ -0,0 +1,35 @@
+{
+ "name": "fold-04-holdout-5",
+ "holdout_indices": [
+ 5
+ ],
+ "holdout_identifiers": [
+ "ess-dive-8ac2940c708a515-20230504T210140482233"
+ ],
+ "exemplar_indices": [
+ 1,
+ 2,
+ 3,
+ 4,
+ 6,
+ 7,
+ 8,
+ 9,
+ 10,
+ 15,
+ 16,
+ 17,
+ 18,
+ 23,
+ 24,
+ 25,
+ 26,
+ 27
+ ],
+ "n_exemplars": 27,
+ "sources": {
+ "package": "data/gold/expert_code/harmonize_sm",
+ "mapping": "data/gold/sm_data_harmonization_mapping.json",
+ "skills": "skills"
+ }
+}
\ No newline at end of file
diff --git a/evaluations/fold-04-holdout-5/agent_outputs/curator_bundle.json b/evaluations/fold-04-holdout-5/agent_outputs/curator_bundle.json
new file mode 100644
index 0000000..5cc6bef
--- /dev/null
+++ b/evaluations/fold-04-holdout-5/agent_outputs/curator_bundle.json
@@ -0,0 +1,272 @@
+{
+ "package_id": "ess-dive-8ac2940c708a515-20230504T210140482233",
+ "doi": "doi:10.15485/1842907",
+ "curator_decision": "INCLUDE",
+ "exclusion_reason": null,
+ "data_payload_files": [
+ {
+ "filename": "SM_SWC.csv",
+ "columns": [
+ "Date Time",
+ "SFA_Location",
+ "Sensor.SN",
+ "Measurement"
+ ],
+ "row_count_estimate": 1089621,
+ "file_size_mb": 49.06,
+ "column_preview": "Date Time,SFA_Location,Sensor.SN,Measurement\n2017-05-26 13:00:00,ER-LMB1R,2012371,0.0896\n2017-05-26 13:00:00,ER-LMB1L,2012373,0.2109\n2017-05-26 14:00:00,ER-LMB1R,2012371,0.0896\n2017-05-26 14:00:00,ER-LMB1L,2012373,0.2124\n2017-05-26 15:00:00,ER-LMB1R,2012371,0.0933",
+ "classification_reason": "Direct soil water content time series with Date Time, SFA_Location, Sensor.SN, and Measurement."
+ }
+ ],
+ "non_payload_files": [
+ {
+ "filename": "SM_AirT.csv",
+ "reason": "Air temperature, not soil moisture."
+ },
+ {
+ "filename": "SM_ST.csv",
+ "reason": "Soil temperature, not soil moisture."
+ },
+ {
+ "filename": "SM_rH.csv",
+ "reason": "Relative humidity, not soil moisture."
+ }
+ ],
+ "location_metadata_files": [
+ {
+ "filename": "Microclimate_Sensors.csv",
+ "columns": [
+ "Logger",
+ "Sensor Type",
+ "Sensor.SN",
+ "SFA_Location",
+ "Site",
+ "Treatment",
+ "Subplot",
+ "Start",
+ "Stop",
+ "Unnamed: 9"
+ ],
+ "row_count_estimate": 130,
+ "content_preview": "Logger,Sensor Type,Sensor.SN,SFA_Location,Site,Treatment,Subplot,Start,Stop,Unnamed: 9\n9735547,Air.T,9721330.0,ER-LMB1,LM,ctl,both,5/26/2017,10/23/2020,\n9735547,rH,9721330.1,ER-LMB1,LM,ctl,both,5/26/2017,10/23/2020,\n9735547,ST,9719166.0,ER-LMB1,LM,ctl,,5/26/2017,10/23/2020,\n9735547,PAR.down,9700885.0,ER-LMB1L,LM,ctl,meadow,5/26/2017,10/23/2020,\n9735547,PAR.up,9700884.0,ER-LMB1L,LM,ctl,meadow,5/26/2017,10/23/2020,"
+ },
+ {
+ "filename": "East_Taylor_Watershed_Community_Observatory_Sites___Point_Locations__Surface_v3_2_20250327.csv",
+ "columns": [
+ "Contact_Name",
+ "Contact_Email",
+ "Contact_Affiliation",
+ "Metadata_Provider_Name",
+ "Location_ID",
+ "Location_Name",
+ "Location_Alias",
+ "Location_Activity",
+ "Location_Category",
+ "Measurement_Categories",
+ "Years_Data_Collected",
+ "Research_Theme",
+ "Description",
+ "Geographic_Region",
+ "Parent_Location_ID",
+ "Latitude",
+ "Longitude",
+ "Lat_Long_Measurement_Method",
+ "Topographical_Elevation",
+ "Topo_Elev_Measurement_Method",
+ "Depth_Profile_Elevations",
+ "Instrumentation",
+ "Variables_Measured",
+ "Related_Reference",
+ "Location_Metadata_Version"
+ ],
+ "row_count_estimate": 1272,
+ "content_preview": "Contact_Name,Contact_Email,Contact_Affiliation,Metadata_Provider_Name,Location_ID,Location_Name,Location_Alias,Location_Activity,Location_Category,Measurement_Categories,Years_Data_Collected,Research_Theme,Description,Geographic_Region,Parent_Location_ID,Latitude,Longitude,Lat_Long_Measurement_Method,Topographical_Elevation,Topo_Elev_Measurement_Method,Depth_Profile_Elevations,Instrumentation,Variables_Measured,Related_Reference,Location_Metadata_Version\nTetsu Tokunaga,tktokunaga@lbl.gov,LBNL,Tetsu Tokunaga,ER-PLM0,Hillslope auger soil samples,,Groundwater Monitoring Well,SFA Intensive,Hydrogeology; Climate,2016,Subsurface water and carbon inventories and fluxes,\"Information needed to quantify water and gas phase flow and transport in the subsurface is being obtained through measurements of profile physical and biogeochemical characteristics, wetting-drying trends, composition of aqueous and gas phases\",PH,PH,38.9195,-106.949646,,2803.56,LIDAR,soils samples down to -0.90m below surface,\"Borehole with Gas Samplers, Water Samplers,Soil Moisture sensors (Matric Potential, TDR), thermistors, piezometer. Click here for drilling log. Click here for instrumentation depths.\",Soil/sediment Temperature; matric potentials; water levels; hydraulic potentials; water content changes; pore water chemical composition; pore gas composition; surface soil CO2 fluxes,,1.0\nTetsu Tokunaga,tktokunaga@lbl.gov,LBNL,Tetsu Tokunaga,ER-PLM1,Hillslope Borehole (M1),,Groundwater Monitoring Well,SFA Intensive,Hydrogeology; Biogeochemistry,Continuing from Dec 2016-Present (as of Dec 2020),Subsurface water and carbon inventories and fluxes,\"Information needed to quantify water and gas phase flow and transport in the subsurface is being obtained through measurements of profile physical and biogeochemical characteristics, wetting-drying trends, composition of aqueous and gas phases\",PH,PH,38.919771,-106.949273,,2787.16,LIDAR,\"-0.30m(PW,PG,T,VWC,matric,EC), -0.58m(PW,PG,T,VWC,matric,EC), -1.05m(PW,PG,T,VWC,matric,EC), -2.05m(PW,PG,T,VWC,matric,EC), -3.60m(PW,PG,T,VWC,matric,EC), -4.72m(PW,T,VWC,matric,EC), -5.40m(PW), -6.89m(PW,piezometer)\",\"Borehole with Gas Samplers, Water Samplers,Soil Moisture sensors (Matric Potential, TDR), thermistors, piezometer. Click here for drilling log. Click here for instrumentation depths.\",Soil/sediment Temperature; matric potentials; water levels; hydraulic potentials; water content changes; pore water chemical composition; pore gas composition; surface soil CO2 fluxes,,1.0\nTetsu Tokunaga,tktokunaga@lbl.gov,LBNL,Tetsu Tokunaga,ER-PLM2,Hillslope Borehole (M2),,Groundwater Monitoring Well,SFA Intensive,Hydrogeology; Biogeochemistry,Continuing from Dec 2016-Present (as of Dec 2020),Subsurface water and carbon inventories and fluxes,\"Information needed to quantify water and gas phase flow and transport in the subsurface is being obtained through measurements of profile physical and biogeochemical characteristics, wetting-drying trends, composition of aqueous and gas phases\",PH,PH,38.920157,-106.948714,,2777.25,LIDAR,\"-0.28m(PW,PG,T,VWC,matric,EC), -0.40m(PW,PG,T,VWC,matric,EC), -0.55m(PW,PG,T,VWC,matric,EC), -1.00m(PW,PG,T,VWC,matric,EC), -1.70m(PW,PG,T,VWC,matric,EC), -2.65m(PW,T,matric), -3.80m(PW,T,VWC,matric,EC), -4.62m(PW,T,matric), -8.20m(PW)\",\"Borehole with Gas Samplers, Water Samplers,Soil Moisture sensors (Matric Potential, TDR), thermistors. Click here for drilling log. Click here for instrumentation depths.\",Soil/sediment Temperature; matric potentials; hydraulic potentials; water content changes; pore water chemical composition; pore gas composition; surface soil CO2 fluxes,,1.0\nTetsu Tokunaga,tktokunaga@lbl.gov,LBNL,Tetsu Tokunaga,ER-PLM3,Hillslope Borehole (M3),,Groundwater Monitoring Well,SFA Intensive,Hydrogeology; Biogeochemistry,Continuing from Dec 2016-Present (as of Dec 2020),Subsurface water and carbon inventories and fluxes,\"Information needed to quantify water and gas phase flow and transport in the subsurface is being obtained through measurements of profile physical and biogeochemical characteristics, wetting-drying trends, composition of aqueous and gas phases\",PH,PH,38.920784,-106.948367,,2759.69,LIDAR,\"-0.20m(PW,PG,T,VWC,matric,EC), -0.41m(PW,PG,T,VWC,matric,EC), -62m(PW,PG,T,VWC,matric,EC), -1.00m(PW,PG,T,VWC,matric,EC), -2.06m(PW,PG,T,VWC,matric,EC), -3.70m(PW,PG,T,matric), -4.85m(PW), -6.50m(PW), -7.80m(PW)\",\"Borehole with Gas Samplers, Water Samplers,Soil Moisture sensors (Matric Potential, TDR), thermistors. Click here for drilling log. Click here for instrumentation depths.\",Soil/sediment Temperature; matric potentials; hydraulic potentials; water content changes; pore water chemical composition; pore gas composition; surface soil CO2 fluxes,,1.0\nTetsu Tokunaga,tktokunaga@lbl.gov,LBNL,Tetsu Tokunaga,ER-PLM4,Hillslope monitoring site (M4),,Groundwater Monitoring Well,SFA Intensive,Hydrogeology; Biogeochemistry,Continuing from Dec 2016-Present (as of Dec 2020),Subsurface water and carbon inventories and fluxes,\"Information needed to quantify water and gas phase flow and transport in the subsurface is being obtained through measurements of profile physical and biogeochemical characteristics, wetting-drying trends, composition of aqueous and gas phases\",PH,PH,38.921006,-106.947953,,2757.43,LIDAR,\"-0.20m(PW,PG,T,VWC,matric,EC), -0.40m(PW,PG,T,VWC,matric,EC), -0.67m(PW,PG,T,VWC,matric,EC), -1.02m(PW), -1.28m(PW)\",\"Shallow Gas Samplers, Water Samplers,Soil Moisture sensors (Matric Potential, TDR), thermistors individually installed (no borehole). Click here for drilling log. Click here for instrumentation depths.\",Soil/sediment Temperature; matric potentials; hydraulic potentials; water content changes; pore water chemical composition; pore gas composition; surface soil CO2 fluxes,,1.0"
+ }
+ ],
+ "sensor_metadata_files": [
+ {
+ "filename": "Microclimate_Sensors.csv",
+ "columns": [
+ "Logger",
+ "Sensor Type",
+ "Sensor.SN",
+ "SFA_Location",
+ "Site",
+ "Treatment",
+ "Subplot",
+ "Start",
+ "Stop",
+ "Unnamed: 9"
+ ],
+ "swc_sensor_rows": 38
+ }
+ ],
+ "readme_content": null,
+ "location_resolution": {
+ "source": "varadharajan_lookup",
+ "qc_flag_recommendation": "g1 for exact Location_ID matches; g2 where no exact Location_ID is available",
+ "site_coordinates": [
+ {
+ "site_id": "ER-LMB1L",
+ "latitude": 38.920264,
+ "longitude": -106.948368,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "ER-LMB1R",
+ "latitude": 38.920272,
+ "longitude": -106.948385,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-LSA1L",
+ "latitude": 38.95598,
+ "longitude": -107.02897,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-LSB1",
+ "latitude": null,
+ "longitude": null,
+ "qc_flag": "g2",
+ "source": "unresolved exact Location_ID in Varadharajan registry raw package"
+ },
+ {
+ "site_id": "WG-LSB1R",
+ "latitude": 38.955732,
+ "longitude": -107.0289,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-LSC1L",
+ "latitude": 38.955129,
+ "longitude": -107.028516,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-UMA2L",
+ "latitude": 38.932854,
+ "longitude": -107.009091,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-UMA2R",
+ "latitude": 38.932865,
+ "longitude": -107.009123,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-UMB2L",
+ "latitude": 38.932668,
+ "longitude": -107.008543,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-UMB2R",
+ "latitude": 38.932673,
+ "longitude": -107.008562,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-UMC2L",
+ "latitude": 38.932596,
+ "longitude": -107.008133,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-UMC2R",
+ "latitude": 38.93261,
+ "longitude": -107.008153,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "WG-USA2R",
+ "latitude": null,
+ "longitude": null,
+ "qc_flag": "g2",
+ "source": "unresolved exact Location_ID in Varadharajan registry raw package"
+ },
+ {
+ "site_id": "WG-USB2R",
+ "latitude": null,
+ "longitude": null,
+ "qc_flag": "g2",
+ "source": "unresolved exact Location_ID in Varadharajan registry raw package"
+ },
+ {
+ "site_id": "WG-USC1L",
+ "latitude": null,
+ "longitude": null,
+ "qc_flag": "g2",
+ "source": "unresolved exact Location_ID in Varadharajan registry raw package"
+ },
+ {
+ "site_id": "WG-USC1R",
+ "latitude": null,
+ "longitude": null,
+ "qc_flag": "g2",
+ "source": "unresolved exact Location_ID in Varadharajan registry raw package"
+ },
+ {
+ "site_id": "XX-ALA1R",
+ "latitude": 38.996955,
+ "longitude": -107.067405,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "XX-ALB1R",
+ "latitude": 38.996908,
+ "longitude": -107.067207,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ },
+ {
+ "site_id": "XX-ALC1R",
+ "latitude": 38.996758,
+ "longitude": -107.067052,
+ "qc_flag": "g1",
+ "source": "Varadharajan location registry raw package"
+ }
+ ]
+ },
+ "time_series_inference": {
+ "is_timeseries": true,
+ "interval_min": "computed per site_id and replicate from timestamp differences; common intervals are 10 or 60 minutes",
+ "reasoning": "Multiple timestamped observations per sensor from 2017-05-26 to 2020-10-23."
+ },
+ "experimental_context": {
+ "manipulation_detected": true,
+ "manipulation_type": "experimental early snowmelt manipulation",
+ "has_control_data": true,
+ "recommendation": "include_control_only",
+ "implementation": "Join SM_SWC.csv to Microclimate_Sensors.csv and retain Treatment == 'ctl'."
+ },
+ "similar_dataset_reference": {
+ "index": 23,
+ "reason": "Fold-local exemplar uses SWC sensor data, joins sensor metadata, filters experimental warming/manipulation records to controls, and computes interval_min per sensor time series."
+ },
+ "open_questions": [
+ "Sensor installation depth is not reported in the available held-out raw files, so depth_m is left missing.",
+ "Five retained control site IDs do not have exact Location_ID matches in the raw location registry and should be treated as g2 for location purposes."
+ ],
+ "summary_counts": {
+ "raw_swc_rows": 1089621,
+ "control_swc_rows_before_datetime_filter": 515825,
+ "raw_swc_locations": 33,
+ "control_swc_locations": 19,
+ "raw_swc_sensors": 37,
+ "control_swc_sensors": 19
+ }
+}
diff --git a/evaluations/fold-04-holdout-5/agent_outputs/harmonize_heldout.py b/evaluations/fold-04-holdout-5/agent_outputs/harmonize_heldout.py
new file mode 100644
index 0000000..bf66a9a
--- /dev/null
+++ b/evaluations/fold-04-holdout-5/agent_outputs/harmonize_heldout.py
@@ -0,0 +1,85 @@
+from __future__ import annotations
+
+from pathlib import Path
+import os
+import numpy as np
+import pandas as pd
+
+PKG_ID = "ess-dive-8ac2940c708a515-20230504T210140482233"
+RAW_DIR = Path(os.environ.get("HELDOUT_RAW_DIR", "/h/jmc/ess-dive_wfsfa_soil_datasets")) / PKG_ID
+OUT_DIR = Path(__file__).resolve().parent
+OUT_CSV = OUT_DIR / "heldout_harmonized.csv"
+
+
+def parse_local_to_utc(series: pd.Series, fmt: str | None, tz: str) -> pd.Series:
+ dt = pd.to_datetime(series, format=fmt, errors="coerce")
+ dt = dt.dt.tz_localize(tz, ambiguous="NaT", nonexistent="shift_forward")
+ return dt.dt.tz_convert("UTC")
+
+
+def ensure_harmonized_cols(df: pd.DataFrame) -> pd.DataFrame:
+ cols = [
+ "datetime_UTC",
+ "site_id",
+ "depth_m",
+ "replicate",
+ "is_timeseries",
+ "interval_min",
+ "volumetric_water_content_m3_m3",
+ "gravimetric_water_content_gH2O_gs",
+ "water_potential_kPa",
+ ]
+ for col in cols:
+ if col not in df.columns:
+ df[col] = np.nan
+ return df[cols]
+
+
+def harmonize() -> pd.DataFrame:
+ swc = pd.read_csv(RAW_DIR / "SM_SWC.csv")
+ sensors = pd.read_csv(RAW_DIR / "Microclimate_Sensors.csv")
+
+ smeta = sensors[sensors["Sensor Type"].astype(str).eq("SWC")].copy()
+ smeta["Sensor.SN"] = pd.to_numeric(smeta["Sensor.SN"], errors="coerce").astype("Int64")
+ smeta = smeta[["Sensor.SN", "SFA_Location", "Site", "Treatment", "Subplot", "Start", "Stop"]]
+
+ x = swc.copy()
+ x["Sensor.SN"] = pd.to_numeric(x["Sensor.SN"], errors="coerce").astype("Int64")
+ x = x.merge(smeta, on=["Sensor.SN", "SFA_Location"], how="left", validate="many_to_one")
+
+ # Remove experimental snowmelt manipulation plots; retain control sensors only.
+ x = x[x["Treatment"].eq("ctl")].copy()
+
+ x["datetime_UTC_dt"] = parse_local_to_utc(x["Date Time"], "%Y-%m-%d %H:%M:%S", "America/Denver")
+ x["site_id"] = x["SFA_Location"].astype(str)
+ x["depth_m"] = np.nan
+
+ sensor_keys = x[["site_id", "Sensor.SN"]].drop_duplicates().sort_values(["site_id", "Sensor.SN"])
+ sensor_keys["replicate"] = sensor_keys.groupby("site_id").cumcount() + 1
+ x = x.merge(sensor_keys, on=["site_id", "Sensor.SN"], how="left", validate="many_to_one")
+
+ x["is_timeseries"] = True
+ x["volumetric_water_content_m3_m3"] = pd.to_numeric(x["Measurement"], errors="coerce")
+ x["gravimetric_water_content_gH2O_gs"] = np.nan
+ x["water_potential_kPa"] = np.nan
+
+ x = x.sort_values(["site_id", "replicate", "datetime_UTC_dt"])
+ x["interval_min"] = x.groupby(["site_id", "replicate"])["datetime_UTC_dt"].diff().dt.total_seconds() / 60.0
+ x.loc[x["interval_min"] <