Claude Code Plugins

Community-maintained marketplace

Feedback

Explore and search the ONT experiment registry with advanced filtering and export capabilities.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name registry-explorer
description Explore and search the ONT experiment registry with advanced filtering and export capabilities.

Registry Explorer Skill

Deep inspection of source data files to extract comprehensive metadata for ONT experiments.

Purpose

This skill explores experiment source data to extract metadata that can't be inferred from names or paths:

  • BAM headers: Basecaller, model, sample info from @RG/@PG tags
  • POD5 files: Run metadata, device info, protocol settings
  • final_summary.txt: Chemistry, basecall model, yield statistics
  • sequencing_summary.txt: Per-read statistics

Quick Start

# Explore a single experiment
/registry-explorer explore exp-abc123

# Scan all local experiments for metadata
/registry-explorer scan --local

# Extract metadata from specific file types
/registry-explorer extract exp-abc123 --bam --pod5 --summary

# Check if source files exist
/registry-explorer check exp-abc123

# List files in experiment directory
/registry-explorer ls exp-abc123

Commands

explore <exp_id> - Full Exploration

Fully explore an experiment's source data:

registry_explorer.py explore exp-abc123 [--verbose]

# Checks for and parses:
# - BAM files (header metadata)
# - POD5 files (run metadata)
# - final_summary.txt (chemistry, model, yield)
# - sequencing_summary.txt (per-read stats)
# - report_*.html (MinKNOW reports)

scan - Batch Scanning

Scan multiple experiments:

registry_explorer.py scan [--local | --public | --all]

# Options:
--missing-chemistry   Only experiments missing chemistry
--missing-model       Only experiments missing basecall_model
--limit N             Process at most N experiments
--apply               Apply changes to registry

extract - Targeted Extraction

Extract specific metadata types:

registry_explorer.py extract exp-abc123 [options]

# Options:
--bam          Extract from BAM headers
--pod5         Extract from POD5 metadata
--summary      Extract from summary files
--all          Extract from all available sources

check - File Existence Check

Verify source files exist:

registry_explorer.py check exp-abc123

# Reports:
# - Directory exists
# - BAM files found
# - POD5 files found
# - Summary files found
# - Total size

ls - List Files

List files in experiment directory:

registry_explorer.py ls exp-abc123 [--recursive]

Metadata Extracted

From BAM Headers

Tag Field Description
@RG SM sample Sample name
@RG PL platform Platform (ONT)
@RG PM model Platform model
@RG DS basecaller Basecaller info
@PG PN program Program name
@PG VN version Program version
@PG CL model Model from command line

From POD5 Files

Field Description
run_id Unique run identifier
experiment_id Experiment name
sample_id Sample identifier
flow_cell_id Flowcell ID
device_id Device serial number
device_type Device model
protocol_group_id Protocol group
protocol_run_id Protocol run ID

From final_summary.txt

Field Description
flowcell_type Chemistry (R10.4.1, R9.4.1)
basecaller Basecaller software
model Basecall model (sup/hac/fast)
yield_bases Total yield in bases
n50 N50 read length
mean_qscore Mean quality score

From sequencing_summary.txt

Field Description
read_count Total reads
pass_count Passing reads
mean_length Mean read length
mean_qscore Mean Q-score

Integration with Registry

Updates to registry include:

  • Metadata fields populated
  • Artifacts tracked (file paths)
  • Provenance updated with extraction timestamp
  • Audit log updated
# Full workflow
/registry-explorer scan --local --missing-chemistry --apply
/registry-scrutinize audit

File Patterns

The explorer looks for files matching:

{location}/
├── *.bam                    # Aligned/basecalled BAM
├── pod5_pass/               # POD5 files
├── final_summary*.txt       # Run summary
├── sequencing_summary*.txt  # Per-read summary
├── report_*.html            # MinKNOW reports
└── other_reports/
    └── *.json               # Additional metadata

Examples

Extract chemistry for all local experiments

/registry-explorer scan --local --missing-chemistry --apply

Deep exploration of single experiment

/registry-explorer explore exp-abc123 --verbose

Check what files exist

/registry-explorer ls exp-abc123 --recursive