A node type cannot be defined outside the kernel assembly: five of Entity's abstract members are internal

首先阅读 Core/Entity/Entity.Definition.cs、Functions/TreeAnalyzer/Sort.Definition.cs、Functions/Evaluation/Evaluation.Definition.cs、Functions/Output/ToSympy.Definition.cs 和 Functions/Continuous/Solvers/InvertNode.Definition.cs 中命名的五个成员定义，以及 Docs/Contributing/Packaging.md 和 Docs/Contributing/Trimming.md。运行 EveryNodeSurvivesEveryPipelineTest 并检查 trimming/AOT gate。完成的标准是：有一份文档化的 contract、一个位于独立 assembly 中的 pipeline test，并保留 reflection 和 AOT 检查。

What is wrong

Entity is public abstract partial record, and its constructor is protected, which reads as an
invitation to derive from it. But a subclass in another assembly does not compile. Five of its
abstract members are internal or private protected, so an external type can never satisfy them:

Deriving from Entity in a separate assembly fails with CS0534, naming each of those five among
the members it cannot implement. This is not a policy that could be relaxed by an
InternalsVisibleTo: a published extension package cannot be on that list.

Why it matters, and what is behind it

This is the concrete blocker under several roadmap items, and it is not the one they name.

#746 item 56 says the extensibility seam
is #338 — looking a type up so it can be
parsed from a string. That is the second half. The first half is being able to write the type at
all, and today you cannot. The same wall sits under item 53 (groups, rings and fields as
first-class, #440) and under the whole
tier-9 domain-package idea, which assumes a geometry or statistics pack can contribute node types.

It also means one thing the packaging decision cannot decide. Docs/Contributing/Packaging.md
(#746 item 78) works out which capabilities belong in the kernel and which ship separately, and its
answer for domain packages is blocked, not decided — because a package boundary is not what is
in the way here. Moving the LaTeX printer or the SymPy exporter out of the kernel runs into the same
five members from the other side: they are abstract members of Entity, so moving them means either
this contract becoming extensible or a visitor over all 68 node types.

Scope

This is a design decision before it is an implementation, and it should be argued before it is
built. The shape of the question:

• Which of the five are genuinely part of a node's public contract, and which are engine internals
  that a visitor or a registry should own instead? ToSymPy and SortHashName look like the second
  kind; Priority and IntrinsicCondition look like the first.
• If a member becomes public or protected, it becomes a compatibility surface — every future
  change to SortLevel, Priority or the InvertNode signature is then breaking. That is the real
  cost and it should be stated before the change, not discovered after.
• Is the answer instead a sealed node hierarchy plus an explicit extension point — a node kind
  that carries a name and an operation table — so that a domain package contributes data rather than
  a subclass? That keeps Entity closed, which is what makes exhaustive matching and the
  EveryNodeSurvivesEveryPipelineTest reflection sweep work today.
• Whatever is chosen must be trimming- and NativeAOT-safe: explicit registration, no runtime assembly
  scanning. See #363,
  #552 and
  Docs/Contributing/Trimming.md.

Acceptance criteria

1. A written decision in Docs/Contributing/, saying what a node type outside the kernel may and may
   not do, and which of the five members are part of the contract.
2. A test that compiles a node type from a separate assembly and puts it through the pipeline —
   parse is not required, but ToString, Latexize, InnerSimplified, substitution and structural
   equality are. A test that only asserts the type exists proves nothing; this has to build.
3. EveryNodeSurvivesEveryPipelineTest still passes, and it is stated whether an external node is
   in scope for it.
4. No new reflection on a hot path, and the trimming/AOT gate stays green.

Dependencies

Independent of everything currently open. It does not depend on
#338 — it is what #338 needs, not the
reverse. It is a prerequisite for #746 items 53 and 56 and for tier 9.

Found while measuring package boundaries for #746 item 78.