adityamparikh opened a new pull request, #272:
URL: https://github.com/apache/solr-mcp/pull/272
Closes #225.
The server instructions already list what the tool surface lacks (deletes,
field drops, `debugQuery`, KNN, non-field facets), but say nothing about
cluster infrastructure. On Kubernetes, a model asked to "restart the bad node"
or "add storage" gets no signal that this server cannot do it and that the Solr
Operator owns it.
## Changes
- `application.properties`: append one sentence to
`spring.ai.mcp.server.instructions`:
> It manages Solr data and configuration only (collections, schema,
documents and aliases). It does not manage cluster infrastructure (nodes,
images, storage, TLS, restarts, scaling); on Kubernetes those are Solr Operator
operations performed with kubectl.
The list names only capabilities that have tools on main. The issue's
draft also listed backups and cluster status; those are #224 and #223, and each
adds its word to this sentence in its own PR.
- `McpClientIntegrationTestBase`: new test asserting that
`getServerInstructions()` from the `initialize` response contains the new
sentence and still contains "Surface boundaries". No existing test asserted the
instruction text. The base class runs under `McpClientIntegrationTest` (http),
`McpClientStdioIntegrationTest` (jar over stdio) and
`DockerImageMcpClientStdioIntegrationTest`.
## Testing
`./gradlew build` passes: 467 tests, 0 failures, 0 skipped. The new test
passed in `McpClientIntegrationTest` and `McpClientStdioIntegrationTest`.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]