Skip to main content
Glama

Servidor MCP SAPUI5

Servidor MCP especializado para desarrollo SAPUI5/Fiori en JavaScript, con herramientas modulares para analisis de proyectos, generacion de codigo, refactorizacion, validacion y consulta de documentacion.

Arquitectura

src/
  index.js
  server/
    mcpServer.js
    toolRegistry.js
  tools/
    agents/
      recommendProjectAgents.js
      materializeRecommendedAgents.js
      scaffoldProjectAgents.js
      validateProjectAgents.js
      saveAgentPack.js
      listAgentPacks.js
      applyAgentPack.js
      refreshProjectContextDocs.js
      recordAgentExecutionFeedback.js
      rankAgentPacks.js
      promoteAgentPack.js
      auditProjectMcpState.js
      upgradeProjectMcp.js
      ensureProjectMcpCurrent.js
      collectLegacyProjectIntake.js
      analyzeLegacyProjectBaseline.js
      buildAiContextIndex.js
      prepareLegacyProjectForAi.js
      scaffoldProjectSkills.js
      validateProjectSkills.js
      recordSkillExecutionFeedback.js
      rankProjectSkills.js
    ui5/
      catalogs/
        ui5ComponentFitRules.js
        ui5SymbolCatalog.js
      generateController.js
      generateFragment.js
      generateFormatter.js
      generateViewLogic.js
      generateFeature.js
      manageI18n.js
      analyzePerformance.js
      analyzeODataMetadata.js
      validateUi5Code.js
      validateUi5VersionCompatibility.js
      validateUi5ODataUsage.js
      scaffoldUi5ODataFeature.js
      securityCheckUi5App.js
    javascript/
      generateFunction.js
      refactorCode.js
      lintCode.js
      securityCheck.js
    project/
      analyzeProject.js
      auditGitWorktreeState.js
      analyzeGitDiff.js
      suggestTestsFromGitDiff.js
      generateCommitMessageFromDiff.js
      prepareSafeCommit.js
      riskReviewFromDiff.js
      generatePrDescription.js
      branchHygieneReport.js
      conflictPrecheck.js
      traceChangeOwnership.js
      smartStageChanges.js
      detectCommitSmells.js
      releaseNotesFromCommits.js
      mergeReadinessReport.js
      mergeActionPlan.js
      readFile.js
      searchFiles.js
      analyzeCurrentFile.js
      syncManifest.js
      writePreview.js
      applyPatch.js
      rollbackPatch.js
      runProjectQualityGate.js
      mcpHealthReport.js
      mcpMetricsDashboard.js
    documentation/
      cacheStore.js
      searchUI5SDK.js
      searchMDN.js
    index.js
  utils/
    fileSystem.js
    manifestSync.js
    parser.js
    refactor.js
    patchWriter.js
    agentPolicy.js
    mcpProjectLayout.js
    xmlParser.js
    validator.js
    logger.js
    telemetry.js
    git.js
    errors.js
    http.js

Related MCP server: ABAP-ADT-API MCP-Server

Herramientas MCP implementadas

  1. analyze_ui5_project

  2. generate_ui5_controller

  3. generate_ui5_fragment

  4. generate_ui5_formatter

  5. generate_ui5_view_logic

  6. generate_ui5_feature

  7. manage_ui5_i18n

  8. analyze_ui5_performance

  9. read_project_file

  10. search_project_files

  11. analyze_current_file

  12. sync_manifest_json

  13. write_project_file_preview

  14. apply_project_patch

  15. rollback_project_patch

  16. run_project_quality_gate

  17. mcp_health_report

  18. tool_contract_catalog

  19. search_ui5_sdk

  20. search_mdn

  21. generate_javascript_function

  22. refactor_javascript_code

  23. lint_javascript_code

  24. security_check_javascript

  25. validate_ui5_code

  26. validate_ui5_version_compatibility

  27. security_check_ui5_app

  28. analyze_odata_metadata

  29. validate_ui5_odata_usage

  30. scaffold_ui5_odata_feature

  31. scaffold_project_agents

  32. validate_project_agents

  33. recommend_project_agents

  34. materialize_recommended_agents

  35. save_agent_pack

  36. list_agent_packs

  37. apply_agent_pack

  38. refresh_project_context_docs

  39. record_agent_execution_feedback

  40. rank_agent_packs

  41. promote_agent_pack

  42. audit_project_mcp_state

  43. upgrade_project_mcp

  44. ensure_project_mcp_current

  45. collect_legacy_project_intake

  46. analyze_legacy_project_baseline

  47. build_ai_context_index

  48. prepare_legacy_project_for_ai

  49. scaffold_project_skills

  50. validate_project_skills

  51. record_skill_execution_feedback

  52. rank_project_skills

  53. audit_git_worktree_state

  54. analyze_git_diff

  55. suggest_tests_from_git_diff

  56. generate_commit_message_from_diff

  57. prepare_safe_commit

  58. risk_review_from_diff

  59. generate_pr_description

  60. branch_hygiene_report

  61. conflict_precheck

  62. trace_change_ownership

  63. smart_stage_changes

  64. detect_commit_smells

  65. release_notes_from_commits

  66. merge_readiness_report

  67. merge_action_plan

  68. mcp_metrics_dashboard

Todas las herramientas se descubren dinamicamente a traves del registro central en src/tools/index.js y se registran en MCP con registerTool(...), incluyendo:

  • name

  • description

  • input schema

  • output schema

Fiabilidad y seguridad

  • Registro de herramientas compatible con JSON-RPC y MCP mediante @modelcontextprotocol/sdk.

  • Validacion estructurada de entrada y salida con zod.

  • Forma de salida determinista mediante structuredContent.

  • Acceso a archivos limitado en sandbox a la raiz del workspace.

  • Proteccion frente a path traversal (bloqueo de .. y de resoluciones fuera de la raiz).

  • Respuestas de error estructuradas con code y message legibles por maquina.

  • Registro centralizado para fallos de herramientas y eventos del ciclo de vida.

  • Aplicacion de politicas a nivel de proyecto mediante .codex/mcp/policies/agent-policy.json en flujos de ranking, recomendacion y puerta de calidad.

  • Verificacion automatica del proyecto MCP al arrancar el servidor (se desactiva con MCP_AUTO_ENSURE_PROJECT=false).

  • Preparacion automatica del contexto legacy al arrancar (se desactiva con MCP_AUTO_PREPARE_CONTEXT=false).

Ejecucion

npm install
npm run start

Observabilidad y logs

El servidor genera telemetria estructurada de uso y rendimiento para analizar la adopcion inicial y detectar cuellos de botella.

  • Eventos por sesion en .mcp-runtime/logs/telemetry-events-<sessionId>.jsonl

  • Resumen agregado en .mcp-runtime/logs/telemetry-session-latest.json

  • Variables de entorno soportadas:

    • MCP_TELEMETRY_ENABLED=false

    • MCP_TELEMETRY_DIR=.mcp-runtime/logs

    • MCP_TELEMETRY_SLOW_THRESHOLD_MS=2000

Ejemplos de llamadas a herramientas

analyze_ui5_project

{
  "tool": "analyze_ui5_project",
  "arguments": {}
}

generate_ui5_controller

{
  "tool": "generate_ui5_controller",
  "arguments": {
    "controllerName": "demo.app.controller.Main",
    "methods": ["onPressSave", "onNavBack"]
  }
}

read_project_file

{
  "tool": "read_project_file",
  "arguments": {
    "path": "webapp/controller/Main.controller.js"
  }
}

generate_javascript_function

{
  "tool": "generate_javascript_function",
  "arguments": {
    "description": "crear un envoltorio fetch con cache",
    "runtime": "node",
    "typescript": false
  }
}

Configuracion MCP para Codex

{
  "mcpServers": {
    "sapui5": {
      "command": "node",
      "args": ["/ruta/absoluta/a/MCPServerUI5/src/index.js"]
    }
  }
}

Documentacion ampliada

Consulta la documentacion para puesta en marcha y mantenimiento en docs/README.md.

Available Tools

51 tools
analyze_current_fileanalyze_current_fileB

Analyze the current file for imports, class structure, and controller methods.

Inputs: path (required). Returns: path, imports, classNames, controllerMethods, +1 more

analyze_legacy_project_baselineanalyze_legacy_project_baselineC

Build a technical baseline for legacy projects to guide high-quality AI integration with minimal context waste.

Inputs: sourceDir, intakePath, baselinePath, baselineDocPath, +5 more. Returns: files, dryRun, intake, changed, +9 more

analyze_odata_metadataanalyze_odata_metadataB

Analyze OData V2/V4 metadata from XML, file, URL, or service root and return entities, properties, navigation, and operations.

Inputs: no named parameters. Returns: model, source, summary, protocol, +1 more

analyze_ui5_performanceanalyze_ui5_performanceC

Analyze UI5 XML/JS files with performance-focused rules and actionable recommendations.

Inputs: sourceDir, maxFiles, maxFindings. Returns: scanned, summary, findings, sourceDir

analyze_ui5_projectanalyze_ui5_projectB

Analyze SAPUI5 project configuration from ui5.yaml, manifest.json, and package.json.

Inputs: no named parameters. Returns: models, routing, namespace, ui5Version, +2 more

apply_agent_packapply_agent_packB

Apply a saved agent pack to the current project with fingerprint verification and scaffolded materialization.

Inputs: no named parameters. Returns: integrity, selectedPack, scaffoldResult

apply_project_patchapply_project_patchC

Apply one or more safe workspace file changes and create rollback metadata for later restore.

Inputs: changes (required), reason. Returns: reason, patchId, appliedAt, changedFiles, +1 more

audit_project_mcp_stateaudit_project_mcp_stateC

Audit MCP project artifacts/layout version and produce a migration plan to the current managed structure.

Inputs: statePath, includeLegacyScan. Returns: state, status, summary, artifacts, +4 more

build_ai_context_indexbuild_ai_context_indexC

Build a token-efficient, quality-aware context index for legacy projects to minimize redundant prompts.

Inputs: sourceDir, baselinePath, intakePath, indexPath, +9 more. Returns: files, dryRun, changed, summary, +5 more

collect_legacy_project_intakecollect_legacy_project_intakeC

Collect and persist legacy-project contextual intake to improve AI guidance quality while minimizing repeated prompts/tokens.

Inputs: intakePath, dryRun, reason, maxDiffLines, +12 more. Returns: dryRun, changed, preview, project, +7 more

ensure_project_mcp_currentensure_project_mcp_currentC

Automatically audit and upgrade MCP project artifacts to the current managed layout when needed.

Inputs: autoApply, force, allowOverwrite, includeVscodeMcp, +12 more. Returns: audit, forced, upgrade, autoApply, +5 more

generate_javascript_functiongenerate_javascript_functionB

Generate an ES2022 JavaScript or TypeScript function with JSDoc for browser or Node runtime.

Inputs: description (required), runtime (required), typescript. Returns: code, runtime, typescript, functionName

generate_ui5_controllergenerate_ui5_controllerB

Generate a SAPUI5 controller skeleton with lifecycle hooks and JSDoc.

Inputs: controllerName (required), methods. Returns: code, controllerName

generate_ui5_featuregenerate_ui5_featureC

Generate an end-to-end UI5 feature scaffold (view/controller/fragment/routing/i18n) with dry-run support.

Inputs: featureName (required), namespace, manifestPath, dryRun, +8 more. Returns: dryRun, changed, feature, previews, +5 more

generate_ui5_formattergenerate_ui5_formatterC

Generate a formatter module that follows SAPUI5 module structure.

Inputs: formatterName, functions. Returns: code, formatterName

generate_ui5_fragmentgenerate_ui5_fragmentC

Generate an XML SAPUI5 fragment skeleton with proper namespaces.

Inputs: fragmentName (required), controls. Returns: code, fragmentName

generate_ui5_view_logicgenerate_ui5_view_logicC

Generate UI5 view-controller event logic methods for XML views.

Inputs: viewName (required), events. Returns: code, viewName

lint_javascript_codelint_javascript_codeB

Run ESLint-style static checks and return warnings plus suggested fixes.

Inputs: code (required). Returns: warnings, suggestedFixes

list_agent_packslist_agent_packsB

List saved reusable agent packs from catalog.

Inputs: packCatalogPath. Returns: packs, exists, packCatalogPath

manage_ui5_i18nmanage_ui5_i18nB

Extract UI literals, detect missing/unused i18n keys, and optionally apply fixes with safe preview/apply patch flow.

Inputs: sourceDir, i18nPath, mode, dryRun, +4 more. Returns: mode, dryRun, changed, summary, +7 more

mcp_health_reportmcp_health_reportB

Return MCP server/runtime health diagnostics (tool exposure, docs alignment, contract snapshot, and managed artifact status).

Inputs: includeToolNames, includeDocChecks, includePolicyStatus, includePolicyTransition, +8 more. Returns: docs, tools, policy, server, +5 more

prepare_legacy_project_for_aiprepare_legacy_project_for_aiC

Prepare a legacy/existing project for high-quality AI delivery by orchestrating ensure, intake, baseline, and context index steps.

Inputs: sourceDir, autoApply, runEnsureProjectMcp, ensureAutoApply, +7 more. Returns: ran, ensure, intake, changed, +6 more

promote_agent_packpromote_agent_packB

Promote/degrade pack lifecycle status (experimental/candidate/recommended/deprecated) using automatic rules over feedback metrics or manual override.

Inputs: no named parameters. Returns: mode, dryRun, changed, preview, +4 more

rank_agent_packsrank_agent_packsC

Rank saved agent packs using execution feedback metrics to prioritize recommendations for the current context.

Inputs: packCatalogPath, metricsPath, policyPath, respectPolicy, +5 more. Returns: exists, policy, summary, metricsPath, +3 more

rank_project_skillsrank_project_skillsC

Rank project skills using local feedback metrics to prioritize reliable skills in future recommendations.

Inputs: catalogPath, metricsPath, minExecutions, maxResults, +4 more. Returns: exists, summary, catalogPath, metricsPath, +1 more

read_project_fileread_project_fileC

Read a workspace file safely with project-root sandboxing.

Inputs: path (required), maxChars. Returns: path, content, truncated

recommend_project_agentsrecommend_project_agentsC

Analyze project signals and recommend agent definitions with materialization-ready arguments.

Inputs: sourceDir, maxFiles, maxRecommendations, includePackCatalog, +18 more. Returns: policy, project, signals, packsMatched, +4 more

record_agent_execution_feedbackrecord_agent_execution_feedbackC

Record agent execution feedback in local JSONL log and update aggregate pack metrics for future ranking and promotion.

Inputs: packSlug (required), projectType (required), outcome (required), packVersion, +16 more. Returns: files, dryRun, record, changed, +3 more

record_skill_execution_feedbackrecord_skill_execution_feedbackC

Record structured execution feedback for project skills and update aggregate metrics for future ranking.

Inputs: skillId (required), outcome (required), qualityGatePass, usefulnessScore, +13 more. Returns: files, dryRun, record, changed, +3 more

refactor_javascript_coderefactor_javascript_codeB

Refactor JavaScript code using modern syntax while preserving readability.

Inputs: code (required). Returns: changes, refactoredCode

refresh_project_context_docsrefresh_project_context_docsC

Refresh docs/mcp context files incrementally using workspace snapshot diff with deterministic output.

Inputs: sourceDir, docsDir, cachePath, maxFiles, +4 more. Returns: delta, dryRun, changed, docsDir, +5 more

rollback_project_patchrollback_project_patchB

Rollback a previously applied patch using stored backup metadata.

Inputs: patchId (required). Returns: patchId, rolledBackAt, restoredFiles, alreadyRolledBack

run_project_quality_gaterun_project_quality_gateB

Run consolidated quality gate for UI5 projects (version compatibility, security, performance, OData usage, and context docs freshness).

Inputs: no named parameters. Returns: pass, checks, policy, reports, +3 more

save_agent_packsave_agent_packC

Save generated agent artifacts into a reusable pack catalog with strict validation and fingerprinting.

Inputs: packName (required), packVersion, blueprintPath, agentsGuidePath, +7 more. Returns: pack, dryRun, changed, previews, +3 more

scaffold_project_agentsscaffold_project_agentsB

Scaffold reusable project agent artifacts (blueprint, guide, bootstrap prompt, optional MCP config) with dry-run and patch safety.

Inputs: projectName, projectType, namespace, outputDir, +12 more. Returns: files, dryRun, changed, project, +4 more

scaffold_project_skillsscaffold_project_skillsC

Scaffold and manage project skills catalog with official references, safe preview/apply flow, and growth-ready feedback artifacts.

Inputs: skillsRootDir, catalogPath, feedbackPath, metricsPath, +8 more. Returns: files, dryRun, changed, project, +3 more

scaffold_ui5_odata_featurescaffold_ui5_odata_featureA

Generate a base UI5 OData feature scaffold (view/controller/manifest/i18n) from service metadata with dry-run safety.

Inputs: no named parameters. Returns: dryRun, changed, feature, metadata, +8 more

search_mdnsearch_mdnC

Search MDN documentation for JavaScript and web platform references.

Inputs: query (required), maxResults, timeoutMs, cache. Returns: query, trace, source, results

search_project_filessearch_project_filesB

Search workspace files by text query with optional extension filtering.

Inputs: query (required), maxResults, extensions. Returns: query, matches

search_ui5_sdksearch_ui5_sdkC

Search official SAPUI5 SDK metadata and return API/topic summaries with examples.

Inputs: query (required), maxResults, timeoutMs, cache. Returns: query, trace, source, results

security_check_javascriptsecurity_check_javascriptA

Detect risky JavaScript patterns including eval, command injection, and prototype pollution.

Inputs: code (required). Returns: safe, findings

security_check_ui5_appsecurity_check_ui5_appA

Scan UI5 XML/JS sources for security risks (XSS, unsafe HTML injection, dynamic code execution, insecure redirects).

Inputs: no named parameters. Returns: safe, scanned, summary, findings, +1 more

sync_manifest_jsonsync_manifest_jsonB

Synchronize manifest.json models, routes, and targets with idempotent updates and optional dry-run.

Inputs: changes (required), manifestPath, dryRun, reason, +1 more. Returns: dryRun, changed, preview, summary, +4 more

upgrade_project_mcpupgrade_project_mcpC

Upgrade MCP-managed project artifacts to the latest layout version with dry-run, preview, and rollback-safe patching.

Inputs: dryRun, allowOverwrite, includeVscodeMcp, runPostValidation, +11 more. Returns: dryRun, changed, previews, migration, +8 more

validate_project_agentsvalidate_project_agentsC

Validate generated agent blueprint and guardrail artifacts for consistency, tool coverage, and MCP integration readiness.

Inputs: blueprintPath, agentsGuidePath, mcpConfigPath, requireMcpConfig, +1 more. Returns: valid, checks, errors, strict, +5 more

validate_project_skillsvalidate_project_skillsC

Validate project skill catalog integrity, official references, and file layout consistency.

Inputs: catalogPath, strict. Returns: valid, checks, errors, strict, +4 more

validate_ui5_codevalidate_ui5_codeC

Validate SAPUI5 code for sap.ui.define usage, dependency order, naming, and MVC separation.

Inputs: code (required), expectedControllerName, sourceType. Returns: issues, isValid, sourceType, issueDetails, +4 more

validate_ui5_odata_usagevalidate_ui5_odata_usageB

Validate UI5 OData usage across manifest, XML/JS bindings, model APIs, and optional metadata cross-checks.

Inputs: no named parameters. Returns: scanned, summary, findings, manifest, +4 more

validate_ui5_version_compatibilityvalidate_ui5_version_compatibilityB

Validate UI5 control/module compatibility against project version and recommend better-fit components (e.g. dedicated date controls).

Inputs: no named parameters. Returns: scanned, summary, findings, references, +3 more

write_project_file_previewwrite_project_file_previewC

Preview a safe workspace file write with hashes, line summary, and textual diff preview.

Inputs: path (required), content (required), maxDiffLines. Returns: path, changed, newHash, oldHash, +6 more

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    C2.9/5.0

    Scored across 51 tools

    Disambiguation3/5

    The tools cover distinct functional areas like analysis, generation, validation, and project management, but there is significant overlap in purpose within categories. For example, multiple analysis tools (analyze_current_file, analyze_ui5_project, analyze_odata_metadata) could confuse an agent about which to use for a specific analysis task, and tools like apply_project_patch and write_project_file_preview both handle file modifications with subtle differences.

    Naming Consistency4/5

    Most tools follow a consistent verb_noun pattern (e.g., analyze_current_file, generate_ui5_controller, validate_ui5_code), which aids readability. However, there are minor deviations such as mcp_health_report (noun_verb) and tools with longer, compound names (e.g., build_ai_context_index) that slightly break the pattern but remain understandable.

    Tool Count2/5

    With 51 tools, the count is excessive for a single server, making it overwhelming and difficult for agents to navigate efficiently. While the domain (SAPUI5 development and project management) is broad, the tools could be better organized or consolidated into fewer, more generalized tools to reduce complexity and improve usability.

    Completeness5/5

    The tool set provides comprehensive coverage for SAPUI5 development, including analysis, code generation, validation, security, performance, project management, and integration with AI/agent workflows. It supports full lifecycle management from legacy project intake to quality gates and patches, with no obvious gaps in the domain's core operations.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      C
      quality
      C
      maintenance
      A Model Context Protocol server that facilitates communication between ABAP systems and MCP clients, providing tools for managing ABAP objects, handling transport requests, and performing code analysis to enhance ABAP development workflows.
      100
      210 npm
      215
      MIT
    • A
      license
      C
      quality
      Not graded
      maintenance
      An MCP server that facilitates seamless interaction with SAP ABAP systems to manage development objects, transport requests, and source code. It provides a comprehensive suite of tools for performing syntax checks, object searches, and code modifications via the ADT API.
      100
      -
    • A
      license
      C
      quality
      D
      maintenance
      An MCP server that enables seamless communication between ABAP systems and MCP clients using the ABAP Development Tools (ADT) API. It provides tools for managing ABAP objects, handling transport requests, and performing code analysis directly through MCP-compatible interfaces.
      100
      MIT