Everything in the atlas

Nothing about it is readable from outside the package: there is no repository, no issue tracker, and the npm page carries no readme, so the README, the examples and the changelog can be read only from the tarball or a CDN mirror, and a bug has nowhere to be filed. Provider URLs are called from the plugin with API keys taken from environment variables in a .env file, so on BTP no call resolves through a destination and the keys live in the environment. The npm line stops at 1.0.1 on 3 November 2025; the changelog shipped inside that version is headed v2.1, so the version npm shows and the version the author counts are not the same.

Quiet Release and repository quiet for 27 months.

Functional boundary Not a test-data factory. It writes files at build time to seed a database, so it does not build fixtures inside a unit test. The declared peer range stops at cds 7 — nothing says it fails on 9, but nothing says it was tried either.

Not published to npm. Install from source.

The maintainer marks this early development; expect the contract to move.

It is a bundle rather than a plugin with one job, and a single release from March 2026 — depending on it for the query helpers brings the auth and validation layers too.

Last released January 2022 at 0.0.21, against a passport strategy Microsoft has since deprecated.

When user-dependent responses are cached without user-aware or explicit operation keys, or translated responses share keys across locales. These cases require key isolation; locale-aware keys are off by default.

Last released April 2022 at 0.0.2, and no repository is declared on the package.

Archived The declared repository is archived.

Functional boundary The repository is archived, so nothing here will move again. Fine to keep where it already runs; anything new should start from what the current cds deploy leaves open rather than from this.

Quiet Release and repository quiet for 25 months.

Functional boundary It performs no authentication flow of its own. The app must already be protected by Entra ID OAuth 2.0 at the App Service level; this only maps the resulting claims. It is explicitly not intended to run locally.

Alerting by mail is not telemetry — if you need traces and metrics correlated across handlers, that is a different layer. It also publishes no repository, so the package is the only thing to read.

No license Neither the package manifest nor the repository tree declares a licence.

Functional boundary The Maven coordinate its README gives, io.github.miyasuta:cds-feature-softdelete:1.0.0, does not resolve on Maven Central as of 2026-09-07 — you build and install it from source.

Last released June 2022 at 0.0.6, and no repository is declared on the package.

Last released July 2022, and no repository is declared on the package.

Explicitly internal utilities, so the surface is not a stable public API and last moved in March 2024. The repository the package declares has been deleted — verified against the GitHub API, not just a failed fetch. The package is still on npm, so it installs; the source it was published from is no longer readable, which is a harder thing to discover than a stale version number.

It declares no peer range on [@sap/cds](artifact:sap-cds) and publishes no repository, so there is nothing to read but the package itself and no CI to look at.

Last released December 2022. Nothing says it fails on current cds, but nothing says it was tried either. The repository the package declares has been deleted — verified against the GitHub API, not just a failed fetch. The package is still on npm, so it installs; the source it was published from is no longer readable, which is a harder thing to discover than a stale version number.

The declared repository is archived.

For a deployed application's UI. The README adds it with `npm add cds-plugin-ui5 -D`, a development dependency of the CDS server, and states nothing about a build or a deployment.

Last released January 2024 at 0.0.6. The repository the package declares has been deleted — verified against the GitHub API, not just a failed fetch. The package is still on npm, so it installs; the source it was published from is no longer readable, which is a harder thing to discover than a stale version number.

It changes what DELETE means for every client of the annotated projection, so anything that must actually remove a row — an erasure request, a retention job — has to go around it.

Release and repository quiet for 46 months.

Built and last touched in the same week of March 2026, so the maintenance reading is what it is — the authentication work is the reason to look, and a stalled repository is the reason to read it before depending on it.

Not reachable Not reachable since 2026-10-20.

Functional boundary The Level A–D grade merges SAP's two published files in the tool's own engine: it is an orientation rather than an ATC result and is left out of the signed audit pack, and the repository documents state by state where the grade is expected to differ from what the clean core ATC check reports. Nothing here runs ABAP Unit — on the RAP track that result is recorded from your own SAP system, and generated CAP tests run against mocks in an isolated runner rather than in a real SAP S/4HANA system. Tenants, organisation accounts, single sign-on, a self-hosted edition and write access through an API are deliberately not built, and one account has five analysis runs before a Gemini API key of its own is needed.

An Eclipse installation older than the 2026-03 release, or a Java below 21, cannot take it: both are listed prerequisites, and 2026-03 is named as the oldest release the current ADT installs on. The tenant, region, client ID and client secret are entered per installation in Eclipse preferences, so each developer configures the Cloud ALM connection in their own IDE.

SAP prohibits gateway/proxy use for SAP Application APIs; the Atlas separately excludes CAP wrappers around those APIs from its assessed scope. Data changes require application-defined unbound actions, and the adapter supplies no rate limiting, agent-action audit logging, approval workflows or policy enforcement.

For deployed environments, the [README](https://raw.githubusercontent.com/cap-java/cds-feature-attachments/302f61c8bb76a0f54f43482ad91d89bf7dc5540f/README.md) excludes the cds-feature-attachments-fs file-system target, which is intended only for local development and testing.

If document versions are required, the [documented restrictions](https://raw.githubusercontent.com/cap-java/sdm/673769f24ef66fd97daa6a6be41262d45065d3fe/README.md) exclude versioned repositories.

Freestyle SQL and `abap_run` mean an agent can read data and execute code, which is well past reading source. Neither is something to grant on a system where the distinction between development tooling and business data matters.

The declared repository is on SAP's internal GitHub (github.tools.sap) and is not publicly reachable, so npm is the only place to read anything about it.

It reads TypeScript exports only; a project written in plain JavaScript has no exported functions for it to extract. A call to the function name in a test file counts as tested, and a call through property access such as `.foo(` does not.

In the Eclipse mode the server lives in a developer's IDE, binds to 127.0.0.1 and answers any request carrying an `Origin` header with HTTP 403, so a pipeline or a remote agent has nothing to reach; the standalone mode on the runtime of the ADT for VS Code extension is the documented route with no IDE open. A new connection starts in write mode, so the 11 write tools are live against a connected system until someone presses Toggle Read-Only. The dropins route for the two bundles is documented as untested.

BUSL-1.1 rather than node-rfc's Apache-2.0, so a product you distribute has to be read against the licence and its Additional Use Grant first, and the source is closed: the GitHub repository is the issue tracker and the documentation, nothing else. `RfcServerOptions.bgRfcHandlers` raises instead of accepting handlers, so a server receiving inbound bgRFC units is outside what this does. An ABAP `MESSAGE`-raised failure still arrives with its T100 template rather than the rendered sentence, and the dump name is not reported as `key` for that case.

Two licences, by which package you install: the standalone server, `@mcp-abap-adt/core`, is AGPL-3.0-only, a licence to read before shipping anything built around it; the library, `@mcp-abap-adt/lib`, is Apache-2.0. The destination model also expects service keys sitting in a local directory, which is a secrets decision to make deliberately.

Release and repository quiet for 33 months.

This is the database, not an application in front of it, so what an agent can read is whatever the connection user can read — the guardrails bound the shape of a query, not the sensitivity of what it returns. Grant the connection the rights you would grant a person.

Its own feature comparison marks connection pooling, automatic reconnect, active-active reads, client-side data encryption, statement distribution, Kerberos authentication, the hdbuserstore, promises and TypeScript types as [@sap/hana-client](https://www.npmjs.com/package/@sap/hana-client) only, so a service that wants a pool or a promise API writes both itself over a callback interface. OpenSSL is its one security provider where the other driver lists SAP CommonCryptoLib, OpenSSL and MSCrypto, and it runs on Node.js 18 or later.

Not for production: the README states local and development instances only, and the server refuses a HAC URL whose host is not private unless `HAC_ALLOW_REMOTE=true` overrides that guard. Masking applies to `config_get` only, so `logs_tail` and `monitoring_info` output, thread dumps included, is unmasked and may carry sensitive data, and whatever a tool returns goes to the MCP client and its model. Certificate checks are disabled whenever `HAC_INSECURE_TLS=true` is set, which the README names as typical for local HAC.

Driving the GUI reaches whatever a user can reach, which is business data and not a development surface. Decide what an agent may do in a session before granting one, not after. Windows only. The HTTP transport is meant for team use but binds to localhost, and unattended multi-user operation is listed as not yet available — a limitation and also the safer default.

LISA does the translation step only: it does not discover objects or manage transports, and its README pairs it with [an ADT MCP server](artifact:sap-adt-mcp-server) for those. Every write needs a transport request, and which object types it can write differs between public cloud or [the SAP BTP ABAP environment](artifact:btp-abap-environment) and on-premise or private cloud, and by release. Running it locally is meant for development and testing, not production.

Development and sandbox systems only: the README warns against pointing it at production. Transport release is never executed, since the tool exists to show the policy stopping it, and separation of duties is checked against the `identity.approvers` list in the policy file rather than against SAP authorisations, so it holds as far as that file does and no further. The write path has been tested with ABAP classes and other object types may need adjustments, there is deliberately no tool that creates or deletes an object, and one server instance serves one SAP system.

Not reachable Not reachable since 2026-10-20.

Functional boundary What it exposes is whatever somebody put in the configuration table, which moves the governance question inside the SAP system and out of the server — an advantage if that table is controlled and a hazard if it is not. The web management interface is a second surface to secure alongside the MCP one.

Its own README opens by calling the server experimental. It also writes: locking, editing and activating in a system an agent can reach is a permissions decision before it is a tooling one.

Every tool its README lists reads — Get and Search, nothing that writes. Locking, editing, activation and transports are the other server, [mcp-abap-abap-adt-api](artifact:mario-andreschak-mcp-abap-adt-api), and mixing the two is how you end up with two MCP servers pointed at one system.

It reads documentation, never a system: nothing in it opens a package.json, a CDS model or an ABAP client, so a question about what your own project does has to be answered elsewhere. The offline half is only as fresh as the submodules the last build pinned, and the index answers where something is written down rather than which of two packages fits a task.

Universal mode trades discoverability for size: one generic tool is harder for a model to pick correctly than a named one, which is why it is not the default. On a small service, standard mode is the better answer.

It fronts APIs you already have; nothing here reads a system's source or model, and a config file is only as good as the entity sets somebody wrote into it. OData V2 is the query surface it names.

The maintainer states 0.x is beta with no production SLA, and scopes support deliberately narrowly: direct application-server transport, classic serialization and password authentication on S/4HANA 2023 and NetWeaver 7.50. Anything outside that is untested rather than unsupported.

Not a system to run a business application on: there is no business application in it, and what it implements of the language, the dictionary and Gateway is a stated subset of each. It is a clean-room reimplementation of the `/IWBEP/` interfaces that bundles no SAP source and no standard DDIC, so a service that reads standard objects has nothing to read, and `ANORMALIES.md` records every place the runtime differs from SAP before any workaround is written.

A requirement that needs code: the skill generates configuration and states that it says so instead of faking it. Dialog and container targets have no verified enter event, so a dashboard placed there loads only through its Refresh button. Production Processes are outside it — it wires PODs to them and does not design them.

Real SAP calls need the proprietary [SAP NetWeaver RFC SDK](external:sap-nw-rfc-sdk) and the `pyrfc==3.3.1` release the README calls archived and yanked, on Python 3.11, so without those only the demo backend answers. Not over HTTP as it stands: credentials are read from the environment for stdio, and the README says not to expose the server over HTTP without implementing the MCP authorization specification and deployment-level rate limiting. Multi-tenant auth and RBAC beyond the one server-wide policy, rate limiting, metrics and tracing, and deployment packaging are listed as not built. The result limits protect the client and not the system, since the README states they cannot reduce work already performed inside SAP, and the MCP tool annotations that describe read and write risk are advisory metadata rather than an authorisation boundary, so the service account's `S_RFC` and table authorisations are still what decides.

It answers about the object catalogue, not about your code: nothing here reads your system, so it will tell you a table is not released and not that you used it. That check is ATC's, and this is what makes ATC's answer actionable.

Ninety tools is a large surface to put in front of a model at once, and client-credentials OAuth means the tenant sees one identity rather than the person asking. Content transport and user management sit in the same tool list as reading a story.

The maintainer states it is unofficial, not endorsed by SAP, and that it uses authenticated Hub web and catalogue endpoints and browser session cookies in ways SAP's terms of service, SAP account terms or an organisation's own policies may restrict. The same README asks the reader not to use it for production, broad distribution or shared access without explicit approval, and says that installing it grants no permission to automate SAP services.

Quiet Release and repository quiet for 30 months.

Functional boundary The annotation sits on a single element, so a rule that spans two fields, or one that needs a lookup, has nowhere in it to go.

Three of its five reference corpora are not shipped: without an api.sap.com key and session cookie, a tenant bearer token and a local SAPUI5 fetch, the REST, MDO and UI5 tools return a message naming the preparation step that is missing. The HTTP transport has no authentication until `MCP_AUTH_TOKEN` is set, so anything past localhost needs that or an authenticating reverse proxy. Packing the ZIP, uploading it through Manage POD 2.0 Extensions and placing the widget in the POD Designer stay manual steps outside the agent, and the SAPUI5 fetch defaults to the SDK's newest version, which the README says is not the version a tenant runs. Node.js 22 or newer.

On ECC the S/4HANA `API_*` services that the three convenience readers default to do not exist, so each one has to be pointed at a Z service of your own or replaced by the RFC tools. `read_table` inherits `RFC_READ_TABLE`'s 512-byte row, so a wide table needs an explicit field list and some fields cannot be selected at all, and the bridge starts a Node process and a fresh logon per call, with no pooling. The audit log is one line per call that includes the call arguments, so every filter and WHERE clause sent to SAP is written to a file that then has to be handled as sensitive.

It needs an S-user with SAP Support Portal access, so there is nothing to search without a support contract behind the login. The maintainer states it uses private SAP APIs behind authentication, asks the reader to check whether that use breaches SAP's terms of service, and disclaims both liability and any guarantee that it keeps working. MFA is handled only by typing the code into a visible browser, so a headless deployment cannot complete a login that demands one.

What it returns is planning material, not a commitment, and the monorepo labels the server's role "future/planned features (planning only)". The maintainer also states it is unofficial, not endorsed by SAP, and that it uses authenticated Road Map Explorer web endpoints and browser session cookies in ways SAP's terms of service or an organisation's policies may restrict, with production, broad distribution and shared access called out as needing explicit approval. An account that requires MFA has to be bootstrapped once through a visible browser.

Neither the package manifest nor the repository tree declares a licence.

No license Neither the package manifest nor the repository tree declares a licence.

Functional boundary The data is illustrative: availability is binary and backup capacity unlimited, and the model carries no inventory; nothing in it is forecast or priced. The recall half runs only against classic ABAP, because its openCypher trace goes through ADBC, which is not released for ABAP Cloud, and it expects an SSH tunnel to the A4H system it was built against. A plain `npm start` passes no model key, so the agent cannot be used on a server started that way.

It is text processing, not an agent: there is no tool calling, no retrieval and no conversation. It needs SAP AI Core, and the backend half needs [@sap-ai-sdk/orchestration](artifact:sap-ai-sdk-orchestration) installed alongside it — an optional peer, so a frontend-only install stays lean.

Neither the package manifest nor the repository tree declares a licence.

Neither the package manifest nor the repository tree declares a licence.

Neither the package manifest nor the repository tree declares a licence.

Neither the package manifest nor the repository tree declares a licence.

No license Neither the package manifest nor the repository tree declares a licence.

Functional boundary Not a descriptor to deploy as it stands: the approuter's `sessionSecret` is written out in `mta.yaml`.

No license Neither the package manifest nor the repository tree declares a licence.

Functional boundary The subaccount destination's url names the us10 trial's repository runtime, so another region or subaccount needs it rewritten before the descriptor deploys.

Neither the package manifest nor the repository tree declares a licence.

It calls the OpenAI API directly with a key in the environment. It was written in early 2023 and last touched in May 2023, which is before SAP AI Core and the CAP LLM plugins existed — so read it for the shape of the application, not for how a CAP service should reach a model today.

Neither the package manifest nor the repository tree declares a licence.

Neither the package manifest nor the repository tree declares a licence.