clio-agent-marketplace

Compare scientific variables with units, ranges, and quality caveats.

Inspect dataset structure before analysis or visualization.

Evaluate data quality from observed structure and summary evidence.

Choose visual checks that match the inspected data.

Route dataset questions to structure, analysis, or visualization experts.

Choose a plot that confirms numeric conversion integrity.

Classify scientific dtype conversions by loss and safety risk.

Coordinate inspect, policy, conversion, integrity, and visual checks.

Apply safe dtype mapping rules for HDF5 to Parquet conversion.

Convert HDF5 tables to Parquet with explicit policy evidence.

Identify source datasets that require conversion policy decisions.

Inspect HDF5 groups, datasets, shapes, and dtypes.

Preserve conversion policy decisions in the final answer.

Decide whether risky fields should be skipped, flagged, or escalated.

Report skipped fields and the policy reason.

Summarize visual checks for converted scientific data.

Recover from unsafe or denied conversion output writes.

Verify missing-value and checksum evidence after conversion.

Verify row counts across source and converted output.

Apply cohort QC threshold reasoning to per-sample variant metrics.

Flag ambiguous sequence or reference regions.

Flag samples with suspicious heterozygosity metrics.

Flag high-impact variants from parsed variant evidence.

Flag per-sample missingness and low call-rate problems.

Identify contig, length, and GC-content evidence in references.

Inspect reference sequence composition before quality claims.

Route cohort variant QC prompts to cohort QC experts.

Route genomics prompts to reference, variant, or cohort experts.

Summarize parsed variant effects and annotation caveats.

Check coordinate bounds and spatial plausibility.

Inspect vector geometry types and feature counts.

Route spatial feature questions through geometry inspection.

Align parsed baseline and candidate Darshan metrics before ranking deltas.

Attribute an HPC I/O regression from aligned metric evidence.

Coordinate baseline/candidate HPC I/O regression analysis through child experts.

Parse Darshan-style text exports into normalized I/O metrics.

Explain performance impact from collective to independent I/O shifts.

Normalize parsed HPC trace counters into comparable summaries.

Parse one baseline Darshan-style trace and preserve evidence for later comparison.

Parse one candidate Darshan-style trace and preserve evidence for later comparison.