发现 Skills

Manage multi-level story synchronization. Use when changes at one abstraction level (pitch, structure, scenes, entities, prose) need to propagate to others, or when story elements feel inconsistent across levels.

Diagnose and calibrate tonal delivery for tabletop RPG sessions. Use when narration feels flat, tone shifts jarringly, descriptions overwhelm play, or energy stays monotonous throughout sessions.

Use when calling TodoWrite with 3+ items, starting multi-step tasks, or when user mentions "board", "kanban", "track progress". Syncs TodoWrite with persistent Kaban board.

Build Next.js 16 applications with modern patterns. Covers Cache Components, proxy.ts, Server/Client Components, Server Actions, and DevTools MCP integration.

This skill should be used when the user asks to "implement LLM-as-judge", "compare model outputs", "create evaluation rubrics", "mitigate evaluation bias", or mentions direct scoring, pairwise comparison, position bias, evaluation pipelines, or automated quality assessment.

Creative task classification skill — accepts a user brief and classifies it as long-form or short-form. Produces classification.json. Tool-agnostic; workflow selection is the caller's responsibility.

Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for submitting diffs or cleanup

Expert in automated QA testing, E2E test automation, visual regression testing, and intelligent test generation. Use when building automated test suites, validating user workflows, testing external integrations, setting up test infrastructure, or implementing AI-powered test discovery.

Query BindingDB for measured drug-target binding affinities (Ki, Kd, IC50, EC50). Search by target (UniProt ID), compound (SMILES/name), or pathogen. Essential for drug discovery, lead optimization, polypharmacology analysis, and structure-activity relationship (SAR) studies.

Interactive visualization library. Use when you need hover info, zoom, pan, or web-embeddable charts. Best for dashboards, exploratory analysis, and presentations. For static publication figures use matplotlib or scientific-visualization.

Applies the operational and scientific frameworks of Stacey Gabriel (genomicist and sequencing platform leader at the Broad Institute). Use this skill whenever the user is designing large-scale biomedical research, operationalizing lab workflows, standardizing data governance, or transitioning from bespoke academic experiments to industrial-scale science. It is highly relevant for queries involving genomics, precision medicine, multiomics, cellular perturbation pipelines, or building massive datasets to fuel AI. Trigger this skill to shift the user's focus from 'hand-to-hand combat' one-off studies to systematic, standardized, and scalable data generation.

Unified Python interface to 40+ bioinformatics services. Use when querying multiple databases (UniProt, KEGG, ChEMBL, Reactome) in a single workflow with a consistent API. Best for cross-database analysis and ID mapping across services. For quick single-database lookups use gget; for sequence/file manipulation use Biopython.

Diffusion-based molecular docking. Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening, for structure-based drug design. Not for affinity prediction.

Parse FCS (Flow Cytometry Standard) files v2.0-3.1. Extract events as NumPy arrays, read metadata/channels, convert to CSV/DataFrame, for flow cytometry data preprocessing.

Medicinal chemistry filters for compound triage. Apply drug-likeness rules (Lipinski, Veber, CNS), structural alert catalogs (PAINS, NIBR, ChEMBL), complexity metrics, and the medchem query language for library filtering.

Bayesian modeling with PyMC. Build hierarchical models, MCMC (NUTS), variational inference, LOO/WAIC comparison, posterior checks, for probabilistic programming and inference.

Create institutional-quality equity research initiation reports through a 5-task workflow. Tasks must be executed individually with verified prerequisites - (1) company research, (2) financial modeling, (3) valuation analysis, (4) chart generation, (5) final report assembly. Each task produces specific deliverables (markdown docs, Excel models, charts, or DOCX reports). Tasks 3-5 have dependencies on earlier tasks.

Organize messy git changes into clean, well-scoped conventional commits. Use when the user wants to clean up their commit history, organize uncommitted changes into logical commits, squash or reorganize recent commits, or curate a changeset before pushing. Triggers on phrases like "clean up my commits", "organize these changes", "split this into commits", "curate history", "make my commits cleaner", or any request to restructure git history into atomic, meaningful units.

.'--- name: delegation-routing description: "Agent-to-agent routing patterns for autonomous delegation via #runSubagent. Defines keyword matching, context templates, model preferences, escalation r...

Test and debug local web applications using Playwright. Captures screenshots, verifies UI behavior, checks browser logs, and runs E2E test scenarios. Use when the user says "test the app", "check the UI", "take a screenshot", "debug the frontend", "run E2E tests", "verify the page works", or "Playwright test".

Observability metrics workflow for metric model design aligned to service health and business impact. Use when teams define or revise service metrics/SLIs for reliable health and capacity decisions; do not use for business-feature implementation logic.

Plan-only workflow for issue/repo changes. Use when user asks to plan, scope, estimate, or design.

Control Home Assistant devices and automations via the REST API. Use this when asked to control lights, thermostats, switches, sensors, media players, automations, or Alexa TTS announcements. Also use it for Home Assistant entity discovery, state queries, and scene/automation management.

Background knowledge for claude-planz workflows. Provides Beads CLI conventions and file structure guidance.