npx claudepluginhub javimontano/jm-adk-alfaThis skill uses the workspace's default tool permissions.
agents/guardian.mdagents/lead.mdagents/specialist.mdagents/support.mdevals/evals.jsonknowledge/body-of-knowledge.mdknowledge/knowledge-graph.mdprompts/meta.mdprompts/primary.mdprompts/variations/deep.mdprompts/variations/quick.mdreferences/functional-tools-patterns.mdtemplates/output.docx.mdtemplates/output.htmlSearches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.
Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.
--- [EXPLICIT] name: functional-toolbelt argument-hint: "module-or-system-name" description: This skill should be used when the user asks to "run event storming", "create a story map", "extract business rules", "write acceptance criteria", or mentions traceability matrix, Given/When/Then, anti-pattern detection, or requirements quality. [EXPLICIT] It provides a 6-tool functional analysis toolkit covering event storming, story mapping, business rule extraction, acceptance criteria writing, traceability matrices, and anti-pattern detection. [EXPLICIT] Use this skill whenever the user needs structured requirements engineering techniques, even if they don't explicitly ask for "functional toolbelt". [EXPLICIT] model: opus context: fork allowed-tools:
Mental models, techniques, and validation tools for producing better functional analysis. NOT a deliverable skill — a toolbelt that enhances output quality of any requirements or specification work. [EXPLICIT]
Las herramientas no producen calidad — la disciplina en su uso sí. El functional toolbelt no es un deliverable — es un kit de técnicas que eleva la calidad de cualquier trabajo de análisis funcional. Event storming sin disciplina es brainstorming caótico. Acceptance criteria sin estructura es prosa ambigua.
$ARGUMENTS format: [tool-name-or-number] [context]
Examples:
"event-storming loan-origination" → Tool 1, domain=loan-origination
"acceptance-criteria user-registration" → Tool 4, story=user-registration
"anti-patterns scan requirements.md" → Tool 6, input=requirements.md
"3 business rules from interviews" → Tool 3, source=interview notes
| Parámetro | Valores | Default | Efecto |
|---|---|---|---|
MODO | piloto-auto, desatendido, supervisado, paso-a-paso | piloto-auto | Nivel de intervención humana durante ejecución de herramienta |
FORMATO | markdown, html, dual | markdown | Formato de salida del artefacto generado |
VARIANTE | ejecutiva, técnica | técnica | Ejecutiva (~40% contenido, resumen visual) vs técnica (full detail + matrices) |
MODO=paso-a-paso → pausa después de cada tool output para revisiónFORMATO=html → aplica design-system tokens al outputVARIANTE=ejecutiva → solo resumen de hallazgos, sin matrices detalladas| FORMATO | Estructura | Uso Principal |
|---|---|---|
markdown | Tablas MD, code blocks para GWT, matrices en texto | Wikis, repos, documentación interna |
html | Styled con design-system tokens, tablas responsive, callouts | Presentaciones a stakeholders, deliverables formales |
dual | Ambos generados; MD como fuente, HTML como render | Cuando el equipo necesita editar (MD) y presentar (HTML) |
Inputs: Domain actors, key processes Process: Identify domain events (temporal markers) > Group by aggregate > Identify commands + actors > Surface policies/rules > Mark hot spots (unknowns) Output: Event timeline > domain events > aggregates > bounded contexts
Edge Cases:
Trade-offs: Speed (90 min) vs completeness (full domain modeling) — prioritize critical paths. Visual (sticky board) vs structured (spreadsheet) — visual aids creativity, spreadsheet aids traceability.
Conditional Logic:
Process: Extract user activities from flows > Decompose to tasks > Write user stories > Assign acceptance criteria > Plot backbone (activities, horizontal) vs detail (stories, vertical) > Slice into releases (MVP, phases 2-3) Output: Backbone > Walking Skeleton (MVP) > Release Plan with story counts
Edge Cases:
Trade-offs: Granularity (tiny stories, many) vs manageability (fewer, less traceable) — target 8-13 stories per release. Detail (full AC now) vs speed (AC during sprint) — write basic AC now, refine during sprint planning.
Conditional Logic:
Techniques: Decision Table (conditions > actions > rows = rules), Decision Tree (condition > branches > leaf = outcome), Rule Catalog (name, classification, condition, action, exception, data, owner)
Classification: Constraint (what cannot happen), Derivation (what is calculated), Inference (what is deduced), Action-Enabling (what triggers a process)
Validation: Completeness (all condition combinations covered?), Consistency (no contradicting rules?), Necessity (traces to business objective?)
Edge Cases:
Trade-offs: Formality (decision tables) vs readability (narrative) — use both. Completeness (all rules) vs speed (80% important rules) — prioritize rules tied to business risk.
Conditional Logic:
Pattern: Given [precondition], When [action], Then [expected result] Variations: Basic (single behavior), Extended (+And), Negative (Then = failure), Boundary (edge values)
Quality Rules (Non-Negotiable):
Anti-Patterns: Vague Then ("handle appropriately"), Missing Given, Implementation leakage ("API endpoint called"), Untestable conditions ("under load" without metric), Conjunctive action (3 outcomes = 3 scenarios)
Edge Cases:
Conditional Logic:
Structure: Requirement ID | Requirement | Linked Use Cases | Linked Flows | Test Cases | AC Covered
Validation Rules:
Coverage Metrics: % traced requirements, % untested flows, % AC without tests
Gap Detection: Requirements without UCs (disconnected), Flows without UCs (orphaned), Orphan test cases (no AC), AC without tests (unvalidated)
Edge Cases:
Conditional Logic:
Categories:
Detection Process: Scan requirements (missing measurements) > Scan AC (implementation details, missing negatives) > Scan rules (untestable conditions) > Scan flows (missing error paths) > Cross-check traceability
Edge Cases:
Conditional Logic:
| Scenario | Adaptation |
|---|---|
| No domain expert available | Use existing documentation + anti-pattern detector to surface gaps; flag as risk |
| Legacy system with no documentation | Start with reverse event storming (trace from outputs back to events) |
| Regulatory domain (healthcare, finance) | Use full 6-tool suite; traceability matrix mandatory with 100% coverage |
| Agile team resisting formal AC | Start with Tool 4 only (GWT); demonstrate value before introducing full suite |
| Microservices with shared bounded contexts | Event storming per service, then cross-reference at integration points |
| Multi-language stakeholders | Standardize domain glossary first (Tool 3 extraction includes terminology) |
| Dimension | Full Suite (6 tools) | Core Set (3-4 tools) | Decision Rule |
|---|---|---|---|
| Coverage | Robust requirements | Acceptable for low-risk | Use full suite for regulated/complex; core for standard |
| Depth vs breadth | Deep on 2 tools | Surface all 6 | Deep dive for complex domains; surface for known domains |
| Facilitation | In-person (high quality) | Async/tool-driven (scalable) | Hybrid recommended |
| Formalism | Decision tables + pseudo-code | Natural language rules | Pseudo-code for critical/financial rules; NL for UI/UX rules |
Before delivering toolbelt output, verify: [EXPLICIT]
functional-spec — consumes toolbelt outputs to build formal specification documentquality-engineering — testing strategy informed by traceability matrix (Tool 5)flow-mapping — process flows feed into event storming (Tool 1) and story mapping (Tool 2)design-system — when FORMATO=html, applies brand tokens to toolbelt outputsPrimary: D-01_Functional_Toolbelt_{project}.md (o .html si {FORMATO}=html|dual) — Functional patterns catalog, composition strategies, implementation guidelines.
Diagramas incluidos:
Author: Javier Montano | Last updated: March 18, 2026
Example invocations: [EXPLICIT]