Use Boltz2 NIM for biomolecular structure prediction and binding affinity. Invoke for Boltz2, protein structures, protein-ligand/DNA/RNA complexes, SMILES or CCD ligands, pIC50/IC50 affinity scoring, mmCIF output, hosted NVIDIA API calls, or local Docker deployment.

Guided pair-programming mode where Claude teaches rather than solves — asking guiding questions, providing hints, explaining concepts, and reviewing user-written code instead of writing solutions. Use when the user wants to learn while building: says "teach me", "help me understand", "guide me", "walk me through", "mentor me", "don't just give me the answer", "learning mode", or invokes /pair.

Explore a codebase and research external patterns before planning a new feature. Spawns parallel subagents for codebase investigation and web research, then synthesizes findings into a compact artifact (~200 lines) for user review. Use when starting a non-trivial feature, investigating unfamiliar code, or evaluating libraries and architectural approaches.

Review a proposed decline of the fictional Tuesday training against the current request, actual calendar, previously sent excuses and previously proposed alternatives. Use for j-claw draft review, without external actions.

Expert en arquitectura completa Brand Studio (post-refactor field-contract-platform): identity, visuals, story, strategy, positioning, narrative (StoryBrand), brand_personality (Jung), communication_assets, contact, team, testimonials, authority_vault. Incluye PersonalityProfile (3-pilar engine) + BuyerPersona (multi-persona). Habla lenguaje de negocio. Use cuando user pide modificar/agregar fields/secciones/buyer-personas/voice-tone/posicionamiento/narrativa/credenciales/equipo/testimonios. Triggers: 'cambiar voz', 'tono de marca', 'arquetipo', 'StoryBrand', 'narrativa', 'posicionamiento', 'UVP', 'unique value proposition', 'buyer persona', 'avatar', 'historia de marca', 'misión', 'visión', 'identidad visual', 'colores marca', 'logo', 'paleta', 'testimonios', 'autoridad', 'credenciales', 'equipo', 'instructor', 'methodology', 'metodología', 'pilares', 'agregar campo brand', 'modificar marca', 'fusionar buyer persona'.

Expert en arquitectura completa Offer Studio (post-refactor field-contract-platform). Cubre fields/secciones/expertise/presets/relaciones/descripciones. Habla lenguaje de negocio. Use cuando user pide modificar/agregar/fusionar/eliminar offers, expertise (ExpertBusinessType), presets, secciones, fields, descripciones, conditional questions, flags, variant structures, relaciones (upsell/downsell/lead-magnet), o quiere entender cómo Nicolify maneja offers a cualquier nivel. Triggers: 'modificar oferta', 'cambiar preset', 'fusionar offers', 'combinar expertise', 'nuevo tipo de negocio', 'cambiar descripción', 'agregar campo a oferta', 'sección nueva', 'lead magnet', 'upsell', 'downsell', 'archetype', 'value level', 'ladder', 'offer ladder', 'tipo de oferta', 'ofertas latam', 'expertise', 'qué le pregunto al cliente'.

Git and GitHub workflow assistant for Nicolify (ap_sales_agent). Creates and manages branches, handles pull requests, resolves merge conflicts, generates changelogs, and manages releases and deployments. Use when the user asks to "create a branch", "sync with github", "push changes", "merge to main", "create a release", "make a PR", "resolve conflicts", "check git status", "deploy to production", "create a version", "write a changelog", or any git/GitHub related task. Also triggers on "quiero hacer un commit", "quiero pushear", "hacemos un release", "pasamos a producción", "nueva versión", "rama nueva".

Full production deployment pipeline: merge development to main, run /test-all fixing all errors, commit, push to trigger GitHub Actions, and monitor the workflow until deployment completes. Use when the user says "pase a producción", "hagamos un pase", "deploy to production", "pasamos a prod", "subamos a producción", or "vamos a producción".

Use when uploaded or project PDF files need exact text extraction, page-specific review, visual inspection of embedded images/diagrams/scans, datasheet figures, tables, or layout-sensitive evidence.

Write or revise hardware product requirements (product/README.md, product/requirements.json, product/contract.json). Use at the start of any hardware product design before parts selection.

Turn an open-ended idea into a PRD plus a runnable definition of done. Use when a developer wants to start a new feature, "file a PRD", "spec out" or "scope" an idea, think through what to build, or kick off the harness for a piece of work. Produces prds/<feature>/prd.md + an executable prds/<feature>/run-prd-test.sh on a prd/<author-slug>/<feature> branch. The one human-attentive skill in the chain.

Scans all authoritative documents (constitution, context, glossary, requirements) and all derivative documents (architecture decisions, architecture overview, MVP architecture spec, vision, implementation heuristics) to add missing citations, replace stale duplicate facts with pointers, and correct contradictions. Use this skill whenever a document has been written or revised in the current phase, whenever you suspect a constraint in one document is not reflected in another, or whenever an agent reading multiple documents has flagged a conflict. Safe to re-run — applying it a second time produces the same result as the first.

Find and compare agent skills across enabled registries, GitHub repositories, search indexes, and local directories; list, enable, or disable those sources. Return broad, deterministic, provenance-rich discovery results, then verify a selected result on demand. Use for discovery and source management, never publishing or silently installing code.

Explains code with visual diagrams and analogies. Use when explaining how code works, teaching about a codebase, or when the user asks "how does this work?", "explain this code", "what does this code do", "walk me through this", "explain this file", "explain this function", "explain this class", "break down this code", or wants to understand code logic, architecture, or behavior. Includes Spring Boot equivalent mappings for non-Java code.

Use when running the retrospective on a piece of work that has just finished — the retro, normally the step straight after `prep-pr` opens the pull request — to write and commit the session-metrics card for the branch, read what the session actually cost against the difficulty declared before it started, and turn that into concrete changes to the skills, `AGENTS.md`, the wiki, the tests or the brief. Also the owner of the card — nothing else writes one carrying the pull request's identity.

Scaffold a new React component following this project's conventions (kel-ui-components first, data-driven content, co-located tests, accessibility by default). Use when asked to create a new component or homepage section.

Access a collection of open-source molecular design and structural biology tools on the Tamarind Bio platform, via its REST API or MCP server — no local GPUs required. Tamarind bundles popular open-source models for structure prediction (AlphaFold, Boltz, Chai, ESMFold), protein, binder, and de novo design (RFdiffusion, ProteinMPNN, BoltzGen), antibody and nanobody design and developability, protein-ligand docking (DiffDock, Autodock Vina), binding-affinity prediction, MSA generation, and molecular dynamics. Use when the user mentions Tamarind or tamarind.bio, wants to run any of these open-source tools in the cloud, references app.tamarind.bio/api or the x-api-key header, or needs to submit batches of sequences for structural or biophysical characterization.

Extract a complete, source-traceable legacy evidence specification for one Oracle Forms module from Forms XML/FMT/FMB/FMX, PLL/PLD, DDL, message catalogues, menus/object libraries, and screenshots. Use for fresh or incremental reverse engineering when the output must preserve every evidenced screen region, tab, field, operation, rule, message, database dependency, provenance locator, uncertainty, and source-supported detail without inventing target requirements, target architecture, target tests, POC assumptions, or implementation decisions.

Run an explicitly requested three-option workflow with three independent executors, one independent paired critic per option, bounded revision loops, and a blind evidence-backed judge. Use only when the user invokes $three-way-judge or explicitly requests this exact panel; this v1 skill does not run the nine-review deep mode.

Run an explicitly requested execution, independent critique, and revision workflow for a task or artifact, with evidence-backed scoring, category floors, a brief security screen, and at most three review rounds. Use only when the user invokes $critic-loop or explicitly requests this bounded critic workflow; do not add it to ordinary reviews.