chore(release): prepare release v2.0.0 (#8)

Bumped versions for: enclave-vm,vectoriadb

Version analysis and docs updates by Codex AI.

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>

Author: github-actions[bot]

CHANGELOG.md

@@ -7,6 +7,15 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.0.0] - 2025-12-12
+
+Transformer-dependent features now load Hugging Face models lazily, with enclave-vm fixing its scorer import path and vectoriadb requiring explicit opt-in for transformer embeddings.
+
+### Updated Libraries
+
+- **enclave-vm** v1.0.2 - LocalLlmScorer lazily imports @huggingface/transformers and treats it as an optional peer dependency.
+- **vectoriadb** v2.0.0 - EmbeddingService now loads transformers dynamically, adds injection hooks, and requires you to install the optional peer when using transformer embeddings.
+
 ## [1.1.0] - 2025-12-11
 
 Hardened ast-guard’s AgentScript preset with additional browser primitive blocks and dynamic import enforcement.

docs/draft/docs/libraries/enclave.mdx

@@ -97,6 +97,12 @@ const enclave = new Enclave({
 
 Leverage the Local LLM scorer to gate executions with on-disk models.
 
+**Note:** `LocalLlmScorer` loads the optional `@huggingface/transformers` peer dependency on demand. Install it before enabling local scoring:
+
+```bash
+npm install @huggingface/transformers
+```
+
 ```typescript
 import { LocalLlmScorer } from 'enclave-vm/scoring';
 

docs/live/docs/libraries/enclave.mdx

@@ -97,6 +97,12 @@ const enclave = new Enclave({
 
 Use the built-in AI scoring gate to evaluate executions with a local model.
 
+**Note:** `LocalLlmScorer` loads the optional `@huggingface/transformers` peer dependency on demand. Install it before enabling local scoring:
+
+```bash
+npm install @huggingface/transformers
+```
+
 ```typescript
 import { LocalLlmScorer } from 'enclave-vm/scoring';
 

libs/enclave-vm/CHANGELOG.md

@@ -7,6 +7,16 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.0.2] - 2025-12-12
+
+### Changed
+
+- Declared @huggingface/transformers as an optional peer dependency so it is only required when local scoring is enabled.
+
+### Fixed
+
+- LocalLlmScorer now imports @huggingface/transformers via a dynamic Function() call, preventing TypeScript and bundler failures when the optional dependency is absent.
+
 ### Changed
 
 - Updated the documented cache directory default for `LocalLlmConfig` to `~/.enclave/models`

libs/enclave-vm/package.json

@@ -1,6 +1,6 @@
 {
   "name": "enclave-vm",
-  "version": "1.0.1",
+  "version": "1.0.2",
   "description": "Sandbox runtime for secure JavaScript code execution",
   "author": "AgentFront <info@agentfront.dev>",
   "homepage": "https://github.com/agentfront/enclave",

libs/vectoriadb/CHANGELOG.md

@@ -7,6 +7,17 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.0.0] - 2025-12-12
+
+### Added
+
+- Added EmbeddingService.setTransformersModule() and clearTransformersModule() to allow injecting custom transformer pipelines (primarily for testing).
+
+### Changed
+
+- Transformer embeddings now lazy-load @huggingface/transformers and emit a ConfigurationError with installation guidance when the package is not installed.
+- @huggingface/transformers is now distributed as an optional peer dependency and must be added explicitly when using transformer embeddings.
+
 ## [1.0.0] - 2025-11-30
 
 ### Added

libs/vectoriadb/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vectoriadb",
-  "version": "1.0.0",
+  "version": "2.0.0",
   "description": "VectoriaDB - A lightweight, production-ready in-memory vector database for semantic search",
   "author": "AgentFront <info@agentfront.dev>",
   "homepage": "https://github.com/agentfront/frontmcp/tree/main/libs/vectoriadb",