seqeron-discovery

name	seqeron-discovery
description	Find the right Seqeron bioinformatics tool or algorithm for a task WITHOUT loading all 427 MCP tool schemas. Use to answer "which Seqeron tool does X?", "is there a tool/algorithm for <biology thing>?", "find the tool for <task>", or before attaching MCP servers / writing a Seqeron C# API call and you need to pick the right tool. Covers melting temperature (Tm), primers, CRISPR guides, alignment (NW/SW/MSA), variant calling, k-mers/assembly, motifs/ORFs, phylogeny, popgen, metagenomics, restriction, codon optimization — across the 11 MCP servers and 247 algorithm docs.
allowed-tools	Bash, Read

Seqeron Discovery

Seqeron exposes 427 MCP tools across 11 servers plus 247 algorithm docs. Attaching every server (or reading every schema) blows up context. This skill routes a plain-language need to the right tool name + Method ID + doc path by grepping the docs — cheap, deterministic, no schemas loaded.

When to use

• The user asks "is there a Seqeron tool/algorithm for X?" or "which tool does X?".
• Before attaching MCP servers or writing a Seqeron.Genomics C# call, to pick the correct tool without loading 427 schemas into context.
• To locate the full I/O schema (follow the doc path) or the C# entry point (Method ID, e.g. PrimerDesigner.CalculateMeltingTemperature).

How to run

python3 scripts/skills/find-tool.py <keywords...> [--server <name>] [--algorithms] [--limit N]

• Keywords are AND, case-insensitive. Matches are ranked: tool name > Method ID > description. Output is a compact table: tool · server · Method ID · doc.
• --server <name> filters to one server dir: core, sequence, parsers, alignment, analysis, annotation, chromosome, metagenomics, moltools, phylogenetics, population.
• --algorithms also searches docs/algorithms/** (title + Test Unit ID + path).
• --limit N caps each section (default 20). --help for usage.

Prefer specific keywords (melting temperature, off target) over broad ones (variant alone returns many hits — add a qualifier or --server).

Examples (verified output)

Melting temperature → Tm tools in sequence + moltools:

$ python3 scripts/skills/find-tool.py melting temperature
TOOL                              SERVER        METHOD ID                                        DOC
melting_temperature               Sequence      SequenceStatistics.CalculateMeltingTemperature   docs/mcp/tools/sequence/melting_temperature.md
primer_melting_temperature        MolTools      PrimerDesigner.CalculateMeltingTemperature       docs/mcp/tools/moltools/primer_melting_temperature.md
primer_melting_temperature_salt   MolTools      PrimerDesigner.CalculateMeltingTemperatureWithSa docs/mcp/tools/moltools/primer_melting_temperature_salt.md

CRISPR → the MolTools guide-design family:

$ python3 scripts/skills/find-tool.py crispr --server moltools --limit 4
TOOL                              SERVER        METHOD ID                                        DOC
crispr_specificity_score          MolTools      CrisprDesigner.CalculateSpecificityScore         docs/mcp/tools/moltools/crispr_specificity_score.md
crispr_system_info                MolTools      CrisprDesigner.GetSystem                         docs/mcp/tools/moltools/crispr_system_info.md
design_guide_rnas                 MolTools      CrisprDesigner.DesignGuideRnas                   docs/mcp/tools/moltools/design_guide_rnas.md
evaluate_guide_rna                MolTools      CrisprDesigner.EvaluateGuideRna                  docs/mcp/tools/moltools/evaluate_guide_rna.md

Alignment, including algorithm docs:

$ python3 scripts/skills/find-tool.py alignment --algorithms --limit 4
TOOL                              SERVER        METHOD ID                                        DOC
alignment_statistics              Alignment     SequenceAligner.CalculateStatistics              docs/mcp/tools/alignment/alignment_statistics.md
format_alignment                  Alignment     SequenceAligner.FormatAlignment                  docs/mcp/tools/alignment/format_alignment.md
...
ALGORITHMS
TITLE                                              TEST UNIT ID          DOC
Global Alignment (Needleman-Wunsch)                ALIGN-GLOBAL-001      docs/algorithms/Alignment/Global_Alignment_Needleman_Wunsch.md
Local Alignment (Smith-Waterman)                   ALIGN-LOCAL-001       docs/algorithms/Alignment/Local_Alignment_Smith_Waterman.md
Multiple Sequence Alignment (MSA)                  ALIGN-MULTI-001       docs/algorithms/Alignment/Multiple_Sequence_Alignment.md

Reading the result

• Doc path (repo-relative) → open it for the full Input/Output schema, errors, and examples. Do not guess the schema; read the doc.
• Method ID → the direct Seqeron.Genomics C# entry point for the API path.
• Server → which MCP server to attach when using the MCP path.
• Test Unit ID (algorithms) → cross-reference into docs/Validation / LIMITATIONS.md for the validated operating envelope.

Point, don't duplicate: this skill routes; the docs are the source of truth.

Fallbacks

• No hits? Loosen keywords (drop qualifiers), try a synonym, or add --algorithms — the concept may be an algorithm doc, not a wrapped MCP tool.
• Optional enrichment: if docs/skills/_generated/catalog.json exists (built by scripts/skills/gen-catalog.py), find-tool.py uses it automatically to annotate results; the script works fully without it.
• Browse a whole server: python3 scripts/skills/find-tool.py <broad-term> --server <name> --limit 100, or ls docs/mcp/tools/<server>/ to list the per-tool docs. Each server also ships a README with its full tool table at src/Seqeron/Mcp/Seqeron.Mcp.<Server>/README.md (Core is src/SuffixTree/Mcp/SuffixTree.Mcp.Core/README.md).
• Once you know the tool, hand off to the relevant domain skill (e.g. bio-moldesign, bio-alignment, bio-annotation) and to bio-rigor for the actual computation.
• To run it, MCP need not be registered anywhere. Call the tool over the shipped MCP server on demand from Python — seqeron-python-client — or via the C# API using the tool's Method ID — seqeron-dev.