SAP × AIForward Deployed Engineer
ENEnglish
中文中EnglishEN한국어KO日本語JA
Talk about your project
WorkServicesOpen SourceAboutContact
中ENKOJA
Open source

Take it, deploy it, change it. Come back when you get stuck.

Three tools live on GitHub: SAP ABAP MCP, SAP CPI MCP and HR Workstation. Four-language READMEs, connection diagnostics, smoke tests and an issue path are included.

Repositories

Three tools, three system boundaries.

The three repositories currently have no LICENSE, so this site says “source-available” rather than formal open source. Learning and internal deployment are fine; contact me before commercial use, redistribution or embedding.

AI assistant to SAP ADT through an MCP server, with design, dev and test subagentsSAP ABAP / MCPLIVEAI ASSISTANTCodex / Claudeintent + toolsMCP SERVERSAP ABAP MCP17 tools + discoverySAP ADTS/4HANArepository + compilerSUBAGENT 01DesignClean Core tiersSUBAGENT 02Developdiagnose + $TMPSUBAGENT 03Testwhere-used + ABAP UnitREAD-ONLY DEFAULTSCOPED WRITESCLEAN CORE01 / 2026
S/4HANA development inside an AI conversation

SAP ABAP MCP

A local MCP server that exposes SAP ADT search, source reading, syntax diagnostics and controlled deployment to AI coding assistants — with design, dev and test subagents guarding Clean Core boundaries.

hand-in-sap-s4hopMCP · SAP S/4HANA · ABAP ADTPublic source · actively maintained
MCPSAP S/4HANAABAP ADTNode.js
Overview
AI agent to SAP Integration Suite through 22 MCP tools, with iFlow and MPL log verificationSAP CPI / MCPLIVEAI AGENTIntent + plandesign / dev / testMCP SERVERSAP CPI MCP22 tools · 17 readSAP BTPIntegration SuiteiFlow · MPL · ODataPACKAGEInventoryread onlyIFLOWSourceedit / deployMPLLogsruntime proofSECURITYMetadatano secretsWRITEGatedallow flagDEPLOY → MPL VERIFYNO SECRET EXPORT02 / 2026
22 tools across the integration lifecycle

SAP CPI MCP

An enterprise MCP server for SAP Cloud Integration: inventory packages and iFlows, export source, deploy, read runtime logs and drill into failures — all inside an AI conversation.

hand-in-sap-integrationsuiteMCP · SAP BTP · Integration SuitePublic source · actively maintained
MCPSAP BTPIntegration SuiteCPI / iFlow
Overview
HR workstation UI to agent plugins, Prisma data and notification channelsHR WORKSTATION / AGENTSLIVEHR USER5 role viewsrecruit → HRDNEXT.JS UIWorkspacetasks + statusAGENT LAYERPlugins + LLMscreening · JD · offerDATAPrisma + SQLiteroster importJOBSState machinePENDING → DONENOTIFYHubWeCom · SMTPRESULTSJSONaudit trail4 CORE MODULESROLE-BASED VIEWS03 / 2026
An AI workspace for HR workflows

HR Workstation

A modular workspace for recruiters, HRBPs, SSC, C&B and HR directors. Plugin agents turn resume screening, JDs, interview guides, offers, automation monitoring and notifications into one traceable pipeline.

hand-in-hcm-workstationNext.js · TypeScript · PrismaPublic source · actively maintained
Next.jsTypeScriptPrismaTailwind
Overview
Deploy yourself

The shortest path

Each README has the full setup, environment variables and troubleshooting. Read it before you start.

hand-in-sap-s4hop

SAP ABAP MCP

Node.js 20+, a reachable SAP S/4HANA ADT endpoint, and an account with ADT permissions.

  1. Clone the repo, run npm install, and follow the README to configure the SAP connection and local MCP client.
  2. Run npm run mcp:doctor for a connection check, then npm run mcp:smoke to verify the protocol and tool surface.
  3. Add the MCP instance to Codex/Claude/Cursor, start read-only, and enable controlled writes only after verification.

Validate in a development client first and use a least-privilege account. Never put production passwords in the repo or prompts.

hand-in-sap-integrationsuite

SAP CPI MCP

Node.js 20+, an SAP BTP Integration Suite tenant, and an OAuth client with the required roles.

  1. Clone, run npm install, copy .env.example and fill in the tenant URL and OAuth client identifiers.
  2. Start with SAP_ALLOW_WRITE=false and use read-only tools to inventory packages, iFlows and runtime state.
  3. Only enable write access when deployment is needed, and keep audit logs plus the MPL verification step.

This tool touches enterprise systems. Use a test tenant or least-privilege client, and never paste credentials, tenant URLs or personal data from logs into an issue.

hand-in-hcm-workstation

HR Workstation

Node.js 20+ and local SQLite are enough; prepare an OpenAI-compatible API config if you want model calls.

  1. Clone, run npm install, copy .env and configure the database plus optional model API.
  2. Run npm run seed for seed data, then npm run dev to open the local workspace.
  3. Walk through a recruitment or automation task, verify notifications and state changes, then connect real data.

Keep demo data separate from real employee data, and confirm privacy and masking requirements before using personal information.

Support

When something breaks, use this path.

Bugs & usage

Open an issue in the repository with version, environment and reproduction steps. Never paste credentials, tenant URLs or personal data from logs.

Open issues

Security & private

Do not post vulnerability details publicly. Use the private form with impact and reproduction conditions; I will reply directly.

Private message

Commercial support

For integration help, customisation, deployment support or an SLA, use the services enquiry and include systems, versions and timing.

Services enquiry
About the licence

The three repositories currently have no LICENSE, so this site says “source-available” rather than formal open source. Learning and internal deployment are fine; contact me before commercial use, redistribution or embedding.

Latest from GitHub

New tools will keep landing here.

SAP ABAP MCP

S/4HANA development inside an AI conversation

hand-in-sap-s4hop2026

SAP CPI MCP

22 tools across the integration lifecycle

hand-in-sap-integrationsuite2026

HR Workstation

An AI workspace for HR workflows

hand-in-hcm-workstation2026

Follow on GitHubVisit the GitHub profile
Keep watching

New tools will keep landing here.

Follow on GitHub, subscribe to repository releases, or choose “open-source support” in the form and I will let you know when a new tool ships.

  • Read-only by default; writes are explicitly gated so AI does not get production access by default.
  • Externalised configuration. Secrets stay out of the repository and out of prompts.
  • Run connection diagnostics and smoke tests before connecting a real system.
Follow on GitHubBrowse repositoriesUsually replies within 24–48 hours
Next step

A system to ship, or a workflow to improve?

Tell me the current state, the outcome you need and your timeline. I reply within 24–48 hours and start with whether it is worth doing — and the fastest path.

Send a private messageEmail directly