Add the get_note tool

Search returns stable IDs. A second tool can now read one complete note without making every search result large.

Add it below search_notes:

  const noteSchema = z.object({
    id: z.string(),
    title: z.string(),
    tags: z.array(z.string()),
    body: z.string()
  })

  server.registerTool(
    'get_note',
    {
      description: 'Read one project note by its exact ID',
      inputSchema: z.object({ id: z.string().trim().min(1).max(100) }),
      outputSchema: z.object({ note: noteSchema }),
      annotations: {
        readOnlyHint: true,
        destructiveHint: false
      }
    },
    async ({ id }) => {
      const note = notes.find(candidate => candidate.id === id)

      if (!note) {
        return {
          content: [{ type: 'text', text: `No note found with ID: ${id}` }],
          isError: true
        }
      }

      const output = { note }
      return {
        content: [{ type: 'text', text: JSON.stringify(output, null, 2) }],
        structuredContent: output
      }
    }
  )

There are two kinds of failure here.

An empty or malformed id fails schema validation before the handler runs. A valid-looking ID that is absent reaches the handler and returns a tool execution error with isError: true.

That distinction gives the model a useful correction path. It can fix the argument shape, or it can search again for a valid ID. Returning an empty success would hide the difference.

Do not include every known ID in the error. On a private server, that would turn one failed lookup into data discovery. The separate search tool is the authorized discovery path.

Test deploy-checklist, missing-note, and an empty string. Confirm that only the successful call contains structuredContent matching the output schema.