AI & ML

holobiomicslab Skill

Use when when a trained model produces probabilistic or ensemble predictions and you need to achieve a specific target accuracy metric (e.g., RMSE ≤ 0.1) or minimize error on a test set, but the unfiltered model does not meet that target.

holobiomicslab Skill

Use when when you have pre-processed MS/MS spectra and a pre-trained Word2Vec model, and need to compute fast, scalable similarity scores for library matching or molecular networking that correlate better with structural similarity than cosine-based methods.

holobiomicslab Skill

Use when when comparing large numbers of MS/MS spectra against spectral libraries or in molecular networking, particularly when molecules differ by multiple structural modifications and cosine-based scores produce excessive false positives.

holobiomicslab Skill

Use when you have a labeled dataset (e.g., mass spectra with molecular structures, SIRIUS 6 fingerprint annotations) that you wish to train a supervised deep learning model on, and you need to estimate generalization performance and reduce variance from a single train–test split.

holobiomicslab Skill

Use when you have a trained regression model (e.g., a neural network or similar predictor) and a held-out test set with ground-truth continuous labels, and you need to measure whether the model's predictions match the true values.

holobiomicslab Skill

Use when when you have trained a predictive model (e.g., neural network or regression model) that outputs continuous scores (such as Spearman correlation coefficients) for individual features (e.

holobiomicslab Skill

Use when you have trained a neural network or regression model on one paired microbiome-metabolome dataset and wish to test whether it can predict metabolite abundances in an independent, externally-sourced dataset collected from different patient cohorts or study populations.

holobiomicslab Skill

Use when when you have preprocessed MS/MS spectral pairs (peak intensities and m/z values) and need to predict molecular structural similarity scores, or when you want to project spectra into a learned chemical embedding space for visualization (e.g., via UMAP) or downstream similarity searches.

holobiomicslab Skill

Use when when you have paired tandem MS spectra and corresponding molecular structures (as SMILES strings or InChI keys) and need to train a model that jointly embeds spectra and structures for structure annotation by database lookup.

holobiomicslab Skill

Use when when you have a trained neural network and need to quantify prediction uncertainty or improve accuracy by filtering low-confidence predictions. Particularly useful when input spectra pairs have variable quality or when downstream tasks (e.

holobiomicslab Skill

Use when during the LDA training phase when you need to decide whether the model has learned a stable representation of Mass2Motifs.

holobiomicslab Skill

Use when you have two LC-MS untargeted metabolomic feature tables (each containing m/z, retention time, and intensity columns) and need to establish which features in dataset A correspond to which features in dataset B, typically for comparative metabolomics, batch effect correction, or.

holobiomicslab Skill

Use when you have a large collection of preprocessed MS/MS spectra (typically >10,000 spectra) with diverse chemical structures and you need to learn embeddings that capture fragmentation patterns and neutral loss relationships.

holobiomicslab Skill

Use when you have preprocessed tandem mass spectrometry spectra converted into a bag-of-fragments representation (with fragments and neutral losses extracted and noise filtered) and your goal is to discover recurring fragmentation patterns or substructures across a large spectral dataset without.

pjt222 Skill

Crea un nuevo archivo de definición de agente siguiendo la plantilla de agente de agent-almanac y las convenciones del registro. Cubre el diseño de la persona, la selección de herramientas, la asignación de habilidades, la elección del modelo, el esquema de frontmatter, las secciones requeridas, la integración en el registro y la verificación de symlinks de descubrimiento. Usar al añadir un nuevo agente especializado a la biblioteca, al definir una persona para un subagente de Claude Code, o al crear un asistente específico de dominio con habilidades y herramientas seleccionadas.

pjt222 Skill

Crea un nuevo archivo SKILL.md siguiendo el estándar abierto Agent Skills (agentskills.io). Cubre el esquema de frontmatter, la estructura de secciones, la escritura de procedimientos efectivos con pares Esperado/En caso de fallo, listas de verificación de validación, referencias cruzadas e integración en el registro. Usar al codificar un procedimiento repetible para agentes, al añadir una nueva capacidad a la biblioteca de habilidades, al convertir una guía o manual de operaciones en formato consumible por agentes, o al estandarizar un flujo de trabajo entre proyectos o equipos.

pjt222 Skill

Evoluciona una definición de agente existente refinando su persona en el lugar o creando una variante avanzada. Cubre la evaluación del agente actual frente a las mejores prácticas, la recopilación de requisitos de evolución, la elección del alcance (refinamiento vs. variante), la aplicación de cambios a las habilidades, herramientas, capacidades y limitaciones, la actualización de metadatos de versión y la sincronización del registro y las referencias cruzadas. Usar cuando la lista de habilidades de un agente está desactualizada, los comentarios de usuarios revelan brechas de capacidad, los requisitos de herramientas han cambiado, se necesita una variante avanzada junto a la original, o el alcance del agente necesita ajuste tras el uso en el mundo real.

pjt222 Skill

agent-almanacエージェントテンプレートとレジストリ規則に従って新しい エージェント定義ファイルを作成する。ペルソナ設計、ツール選択、スキル 割り当て、モデル選択、フロントマタースキーマ、必須セクション、レジストリ 統合、発見シンリンクの確認をカバーする。新しい専門エージェントをライブラリ に追加する場合、Claude Codeサブエージェントのペルソナを定義する場合、 またはキュレートされたスキルとツールを持つドメイン固有のアシスタントを 作成する場合に使用する。

pjt222 Skill

Agent Skills オープン標準（agentskills.io）に従って新しいSKILL.mdファイルを 作成する。フロントマタースキーマ、セクション構造、Expected/On failureペアを 持つ効果的な手順の作成、バリデーションチェックリスト、クロスリファレンス、 レジストリ統合をカバーする。繰り返し可能な手順をエージェント用に体系化する 場合、スキルライブラリに新しい機能を追加する場合、ガイドやランブックを エージェント消費可能な形式に変換する場合、またはプロジェクトやチーム間で ワークフローを標準化する場合に使用する。

pjt222 Skill

既存のエージェント定義をそのペルソナをその場で改良するか、上級バリアントを 作成することで進化させる。現在のエージェントのベストプラクティスに対する 評価、進化要件の収集、スコープ（改良 vs バリアント）の選択、スキル、 ツール、機能、制限事項への変更の適用、バージョンメタデータの更新、 レジストリとクロスリファレンスの同期をカバーする。エージェントのスキル リストが古くなっている場合、ユーザーフィードバックが機能のギャップを 明らかにした場合、ツール要件が変わった場合、元のエージェントと並んで 上級バリアントが必要な場合、または実際の使用後にエージェントのスコープを 調整する必要がある場合に使用する。

holobiomicslab Skill

Use when you have raw MS2 spectra (m/z and intensity pairs) that you want to match against a large training dataset of annotated library spectra (e.g., GNPS), and you need to reduce noise and computational burden before applying kernel-based scoring methods such as IOKR.

holobiomicslab Skill

Use when you have discovered Mass2Motifs or other fragmentation pattern representations via LDA and need to generate vector embeddings to query a reference motif database (MotifDB) for structural annotation candidates.

holobiomicslab Skill

Use when you have acquired targeted mass spectrometry data with measured ion intensities for known standard compounds at multiple concentration levels, and you need to convert sample intensities into absolute or relative concentrations.

holobiomicslab Skill

Use when when you have a high-resolution mass spectrometry dataset with m/z values and need to generate Kendrick mass plots where users should choose between plotting raw m/z or computed Normalized Kendrick Mass (NKM) on the x-axis.

holobiomicslab Skill

Use when you have a trained formula ranking model (such as MIST-CF) and want to measure the specific performance gain from incorporating multiple positive-mode adduct types (e.g., [M+H]+, [M+Na]+, [M+K]+, [M+NH4]+) instead of restricting predictions to [M+H]+ only.

holobiomicslab Skill

Use when when you need to understand how a multi-instrument mass spectrometry platform (like mzmine) decides which processing module receives a given dataset based on its declared data type (LC vs. GC vs. IMS vs. MS imaging).

holobiomicslab Skill

Use when you have measured intracellular metabolite abundances (LC-MS or similar) across multiple cell lines or conditions and a stoichiometric metabolic model (with reaction-metabolite associations) to estimate how differences in substrate availability—independent of gene expression—translate into.

holobiomicslab Skill

Use when when deploying a Word2Vec-based spectral similarity model (such as Spec2Vec) on a new mass spectrometry dataset and needing to assess whether the pre-trained model's learned peak embeddings sufficiently represent the peaks in your query spectra.

holobiomicslab Skill

Use when you have developed a predictive model and need to compare its performance against established baselines (e.g., linear regression, Random Forest, Canonical Correlation Analysis) across multiple datasets with paired input-output features.

holobiomicslab Skill

Use when you have two peak-picked, conventionally aligned LC-MS metabolomics datasets (e.

holobiomicslab Skill

Use when you have a working base MPNN model (e.g., chemprop) and need to add task-specific feature processing layers (spectral, electronic, or domain features) to improve predictions on a specialized molecular property or spectrum.

holobiomicslab Skill

Use when you have two embedding tensors of identical shape (e.g., both 512-dimensional) and need to produce a fused representation that captures multiplicative interactions between modalities.

holobiomicslab Skill

Use when you have a generic genome-scale metabolic model (SBML format) and cross-sectional omics data (RNA-seq, intracellular metabolomics, extracellular flux measurements from bioanalyzer or similar) from multiple biological samples (cell lines, conditions).

holobiomicslab Skill

Use when when you need to establish comparable performance baselines for a novel spectrum prediction model and require fair comparison across multiple baseline architectures. Trigger this skill when: (1) you have a new spectrum prediction approach (e.g., ICEBERG, SCARF) to benchmark;

holobiomicslab Skill

Use when you have tunemix or other reference standards with known m/z, drift-time, charge state, and CCS values, and you need to establish a predictive calibration model for your ion-mobility mass spectrometry instrument.

holobiomicslab Skill

Use when you have ESI/LC-MS test spectra requiring candidate metabolite ranking, pre-trained MLP (NEIMS) and GNN baseline models are available or can be trained, you seek quantified improvement over single-model average rank performance (baseline MLP shows ~339 average rank), and your evaluation.