Release v0.3.8: bump package + plugin to 0.3.8

.claude-plugin/marketplace.json

@@ -7,7 +7,7 @@
   },
   "metadata": {
     "description": "The trust layer between AI and your data. A governed semantic model for AI-to-database access — every join approved, every metric signed off, every answer with a receipt. Local-first: no infra, no servers, no data leaves your machine.",
-    "version": "0.3.7",
+    "version": "0.3.8",
     "pluginRoot": "./plugins",
     "tags": ["data", "sql", "governance", "trust-layer", "semantic-model", "local-first", "fair-code"],
     "repository": "https://github.com/AgamiAI/agami-core",
@@ -18,7 +18,7 @@
       "name": "agami-core",
       "source": "./plugins/agami",
       "description": "A governed trust layer between AI and your database: every join FK-derived or human-approved, every metric signed off, every answer with a receipt (the SQL, the model version, who vouched for each definition). Runs entirely on your machine via Claude's Bash / Read / Write tools — no MCP server or Python required if you have psql/mysql or DuckDB (an optional local MCP server is available for use from Claude Desktop).",
-      "version": "0.3.7",
+      "version": "0.3.8",
       "keywords": ["database", "governance", "trust-layer", "semantic-model", "sql", "postgres", "snowflake"],
       "category": "data"
     }

CHANGELOG.md

@@ -10,7 +10,7 @@ is the source of truth a host installs against — bumping it is what invalidate
 user's plugin cache (see [CONTRIBUTING.md](CONTRIBUTING.md)). Each released section
 below corresponds to one such version.
 
-## [Unreleased]
+## [0.3.8] — 2026-07-06
 
 ### Added
 
@@ -238,6 +238,7 @@ First version tracked in this changelog. Earlier history lives in the git log.
   other clients — stdio, no auth, no network.
 - Fan-trap / chasm-trap pre-flight that refuses to silently double-count.
 
+[0.3.8]: https://github.com/AgamiAI/agami-core/compare/v0.3.7...v0.3.8
 [0.3.7]: https://github.com/AgamiAI/agami-core/compare/v0.3.6...v0.3.7
 [0.3.6]: https://github.com/AgamiAI/agami-core/compare/v0.3.5...v0.3.6
 [0.3.5]: https://github.com/AgamiAI/agami-core/compare/v0.3.4...v0.3.5

packages/agami-core/pyproject.toml

@@ -11,7 +11,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "agami-core"
-version = "0.3.7"
+version = "0.3.8"
 description = "agami-core — governed semantic model, the shared MCP TOOLS harness, and the unified local query executor."
 readme = "README.md"
 requires-python = ">=3.10"

plugins/agami/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "agami-core",
   "description": "The trust layer between AI and your database, so an agent's answers are defensible. Every join, metric, and named filter is confidence-scored and either FK-derived or human-approved via a review dashboard. Every answer ships a SQL receipt showing what ran and which model version it used. Works across Postgres / MySQL / Snowflake / BigQuery / Redshift / SQL Server / Oracle / Databricks / Trino / DuckDB / SQLite. Local-first: credentials, schema, results never leave your machine.",
-  "version": "0.3.7",
+  "version": "0.3.8",
   "author": {
     "name": "Agami AI",
     "email": "skills@agami.ai",