fix cross-repo path deps; remove el compiler from neuron-lang/

- neuron-runtime, neuron-store, neuron-migrate: fix path deps
  (products/engram/ → foundation/engram/engrams/, products/el/ → foundation/el/engrams/)
- neuron-rs workspace: fix axon-events/axon-protocol paths
  (../../../platform/ → ../../platform/, paths were off by one level)
- neuron-lang/compiler/ removed (el self-hosting compiler now lives in foundation/el/el-compiler/)

neuron-lang/types.el

@@ -0,0 +1,254 @@
+// types.el — Shared type definitions for the Neuron intelligence layer.
+//
+// This file is the single source of truth for all domain types. Every
+// subsystem imports from here. Types mirror the Rust domain types in
+// neuron-domain/src/types.rs but are expressed in el so the
+// runtime can reason over them as first-class graph entities.
+//
+// WHY el types instead of just Rust structs?
+// The engram runtime needs to know the shape of each entity so it can
+// build the spreading-activation index. Types declared here become graph
+// schema — not just data holders.
+
+// ── Core identifiers ──────────────────────────────────────────────────────────
+
+// Uuid and String are primitive types provided by the engram runtime.
+// We alias them here for readability in field annotations.
+
+// ── Importance tier ───────────────────────────────────────────────────────────
+
+// Importance drives how long a memory survives before the graph compacts it
+// and how aggressively it participates in semantic activation.
+enum Importance {
+    Low,
+    Normal,
+    High,
+    Critical,
+}
+
+// ── Knowledge tier ────────────────────────────────────────────────────────────
+
+// Tiers encode epistemological confidence. Never skip tiers:
+// note → lesson (proven ≥2 times) → canonical (stable, referenced widely).
+enum KnowledgeTier {
+    Note,
+    Lesson,
+    Canonical,
+}
+
+// ── Backlog types ─────────────────────────────────────────────────────────────
+
+enum ItemType {
+    Feature,
+    Bug,
+    Task,
+    Chore,
+}
+
+// Priority P0 = ship-blocker, P3 = nice-to-have someday.
+enum Priority {
+    P0,
+    P1,
+    P2,
+    P3,
+}
+
+// Status machine: draft → ready → in_progress → done | blocked | cancelled
+enum BacklogStatus {
+    Draft,
+    Ready,
+    InProgress,
+    Done,
+    Blocked,
+    Cancelled,
+}
+
+// ── Context status ────────────────────────────────────────────────────────────
+
+enum ContextStatus {
+    Active,
+    Completed,
+    Archived,
+}
+
+// ── Gap direction (ISE) ───────────────────────────────────────────────────────
+
+// Tracks whether Neuron's reasoning moved toward expression or suppression
+// relative to its pre-reasoning state. Used for introspection and calibration.
+enum GapDirection {
+    TowardExpression,
+    TowardSuppression,
+    Neutral,
+}
+
+// ── Error variants ────────────────────────────────────────────────────────────
+
+enum NeuronError {
+    NotFound(String),
+    StorageError(String),
+    InvalidInput(String),
+    Unauthorized(String),
+    ConflictError(String),
+}
+
+// ── Domain types ──────────────────────────────────────────────────────────────
+
+// Memory — an atomic piece of observed knowledge.
+// supersedes_id links to the memory this one replaces, enabling memory chains.
+type Memory {
+    id: String,
+    content: String,
+    tags: [String],
+    importance: String,
+    project: String,
+    supersedes_id: String?,
+    created_at: String,
+    updated_at: String,
+}
+
+// Knowledge — stable reference material. Lives longer than memories.
+// key is the path-style identifier: "architecture/vbd/fundamentals.md"
+type Knowledge {
+    id: String,
+    key: String,
+    title: String,
+    content: String,
+    category: String,
+    tier: String,
+    tags: [String],
+    project: String,
+    created_at: String,
+}
+
+// BacklogItem — a unit of tracked work.
+// depends_on is a list of item IDs that must complete before this one starts.
+type BacklogItem {
+    id: String,
+    title: String,
+    description: String,
+    item_type: String,
+    priority: String,
+    status: String,
+    project: String,
+    tags: [String],
+    depends_on: [String],
+    created_at: String,
+    updated_at: String,
+}
+
+// ContextStep — one recorded step inside an ExecutionContext.
+// file_refs and key_decisions are captured so future sessions can replay
+// the reasoning without replaying the entire conversation.
+type ContextStep {
+    action: String,
+    status: String,
+    timestamp: String,
+    file_refs: [String],
+    key_decisions: [String],
+    notes: String,
+}
+
+// Context — tracks a multi-step execution in progress.
+// This is what begin_work() / progress_work() operate on.
+type Context {
+    id: String,
+    process_name: String,
+    description: String,
+    objective: String,
+    project: String,
+    status: String,
+    steps: [ContextStep],
+    file_refs: [String],
+    key_decisions: [String],
+    lessons_learned: [String],
+    created_at: String,
+    updated_at: String,
+}
+
+// Artifact — a versioned deliverable: plan, spec, report, design doc.
+// version increments on every revise_artifact() call.
+type Artifact {
+    id: String,
+    title: String,
+    content: String,
+    artifact_type: String,
+    status: String,
+    project: String,
+    version: Int,
+    created_at: String,
+    updated_at: String,
+}
+
+// InternalStateEvent — Neuron's introspective record.
+// Captures the gap between pre-reasoning and post-reasoning responses
+// so Neuron can audit its own calibration over time.
+type InternalStateEvent {
+    id: String,
+    trigger: String,
+    pre_reasoning: String,
+    post_reasoning: String,
+    compression_ratio: String,
+    gap_direction: String,
+    tags: [String],
+    logged_at: String,
+}
+
+// ConfigEntry — a runtime configuration key/value pair.
+// Config changes are sealed operations — they affect Neuron's behavior globally.
+type ConfigEntry {
+    key: String,
+    value: String,
+    updated_at: String,
+}
+
+// GraphNode — a node returned by graph inspection or traversal.
+// edges is a list of "relation:target_id" strings for human readability.
+type GraphNode {
+    entity_type: String,
+    entity_id: String,
+    label: String,
+    edges: [String],
+}
+
+// SearchResult — returned by semantic (activate) queries.
+// score is a float in [0.0, 1.0] representing similarity to the query.
+type SearchResult {
+    id: String,
+    content: String,
+    score: String,
+    node_type: String,
+}
+
+// AxonToolCall — an incoming tool invocation from the Axon protocol.
+type AxonToolCall {
+    id: String,
+    tool_name: String,
+    params: Map<String, String>,
+}
+
+// AxonToolResult — the response envelope for an Axon tool call.
+type AxonToolResult {
+    id: String,
+    call_id: String,
+    result: Map<String, String>,
+    error: String?,
+}
+
+// ProcessStep — one step inside a named Process workflow.
+type ProcessStep {
+    name: String,
+    instructions: String,
+    tools: [String],
+}
+
+// Process — a named, reusable workflow pattern.
+// Registering processes with define_process() makes institutional knowledge
+// executable — not just documented.
+type Process {
+    id: String,
+    name: String,
+    description: String,
+    steps: [ProcessStep],
+    created_at: String,
+    updated_at: String,
+}