onedevatlasBeta
Browse the atlas
Person

Yassin Bahri

2maintains
GitHub@yassinbahriWebsiteyassinbahri.github.iolast push

Capabilities4

Provides 4
Agent access to your codeABAPBTPAgent access to your codecapabilityLetting a coding agent read the project or the system you actually have, reach its model, its objects and its checks, and act on it. In file-based environments that works out of the box; complex systems need dedicated access capabilities. Whatever the environment, the agent also needs project- and framework-specific context, so it understands the thing in front of it rather than guessing from what it remembers of its training data. The transport is incidental: an MCP server, a language server or a CLI answer the same question, and what differs between them is how much of the real thing the agent can see. Distinct from [MCP servers](capability:mcp-servers), which is a protocol adapter on your own running service, and from [agent guidance](capability:agent-guidance), which is knowledge handed over as files rather than read from the thing itself.Agent access to your systemsBTPABAPAgent access to your systemscapabilityGiving an AI agent access to the business data, configuration or operations of systems you run through an MCP server. [Serving your application over MCP](capability:mcp-servers) exposes your own service contract; system access reaches the systems behind it, and one implementation may do both. Development assistance belongs to [agent access to your code](capability:agent-code-access).Retrieval-augmented generationABAPBTPRetrieval-augmented generationcapabilityGrounding a model's answer in data retrieved at question time: text split into chunks, each chunk embedded as a vector, and the chunks nearest to the question placed into the prompt. Semantic search is the same retrieval with a person reading the matches, and fine-tuning changes the model's weights, not its input.Tool governanceABAPBTPTool governancecapabilityControlling what an agent may call, with which authorisations, and leaving an audit trail of what it did. The authorisation question that MCP exposure raises but does not answer.

Entries2

Group by
EntrySignals
Maintains2
pyrfc-mcp
Functional boundary

MCP server calling RFC-enabled function modules behind an allowlist

sap-knowledge-pipeline
Functional boundary

Python library turning OData or HANA records into chunked documents

Details

History4
  • pyrfc-mcpprovides → capability:agent-systems-accessThe concern it answers is reaching into an SAP system from outside, split out of MCP servers on this day; the sentence on the edge is unchanged.
  • pyrfc-mcpentry → createdHanded over as a URL. Filed behind the ABAP door because the system it reaches is an ABAP one and the reader who points it at theirs is working on that system, which is the opposite case from [open-rfc](artifact:marianfoo-open-rfc), a client an ABAP developer would never install. Three edges: [MCP servers](capability:mcp-servers) for the tool surface over function modules, [agent access to your code](capability:agent-code-access) for `read_abap_source` alone, and [tool governance](capability:tool-governance) for the allowlists, the table guard, the result limits and the audit log, which the README states as the point of the server rather than as a footnote. [RFC connectivity](capability:rfc-connectivity) was considered and left off: the RFC client here is [PyRFC](external:pyrfc), so nobody installs this to call a function module from their own code. No flavor is stated, since the README names no ABAP release or cloud edition, and `auth` and `tool_count` are left out because the README names neither a mechanism nor a number.
  • sap-knowledge-pipelineentry → createdHanded over as a URL. Filed behind the BTP door because what it reads is an SAP OData service and [SAP HANA](artifact:sap-hana), and neither the library nor its recipes touch CAP, UI5 or ABAP; the repository's own topics are rag, rag-pipeline, sap, sap-btp and sap-hana. One edge, [retrieval-augmented generation](capability:retrieval-augmented-generation): the capability's CAP coverage says the layer above the vector, splitting long text into chunks and deciding what goes into the prompt, is hand-written in every entry there, and that layer is what this library is. [Data import and export](capability:data-import-export) was considered and left off, since that capability is about a file somebody supplies rather than records read out of a live service, and [OData](capability:odata) is about serving a model as OData rather than consuming one. The optional SAP `hdbcli` driver is named in the prose and carries no edge: SAP's Node.js driver is the artifact [@sap/hana-client](artifact:sap-hana-client), so its Python sibling belongs in the graph the same way and not as an external, which is an editorial decision rather than one for this batch.
  • First entry seenDiscovered via Added by hand.
Profile1
For your agent
Ref
actor:yassinbahri
JSON
/actors/yassinbahri.json
MCP
Connecting an agent