msty-admin-mcp 6.2.0__tar.gz → 6.2.2__tar.gz

msty_admin_mcp-6.2.2/.dockerignore

@@ -0,0 +1,23 @@
+.scratch
+.venv
+venv
+dist
+build
+.pytest_cache
+__pycache__
+*.pyc
+.git
+.eggs
+*.egg-info
+.coverage
+htmlcov
+tests
+.DS_Store
+.env
+.env.local
+*.db
+*.sqlite
+*.sqlite3
+*.bak
+*.pem
+*.key

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/CHANGELOG.md

@@ -2,6 +2,66 @@
 
 ## Unreleased
 
+## v6.2.2 - 2026-11-10
+
+Safety release for Msty Studio 3.0.
+
+### Added
+
+- Studio version gate. The app version is parsed as numbers, so 10.0.0 sorts above 2.9.12. Studio 2.4 to 2.9.x is supported. Versions below 2.4 and from 3.0 on are unsupported. A missing, unreadable or over-long version is unknown.
+- `get_msty_installation` reports `layout` and `layout_status`.
+- Studio read results carry a warning when the version is unknown, because the data may be a stale copy.
+- macOS detection also looks for `~/Applications/MstyStudio.app`.
+
+### Changed
+
+- On Studio 3.x and below 2.4 every Studio storage read and write is refused with one clear message. Detection no longer falls back to the legacy `msty.db` files or a search of the `File System` folder, because after a 3.0 migration those may be stale copies.
+- Persona and prompt writes need a detected Studio 2.4 to 2.9.x. They are refused on other versions and when the version is unknown, and an explicit database path cannot bypass this. The JSON pack is still saved for import in Persona Studio.
+- On Windows and Linux the Studio version cannot be read yet, so Studio writes are refused there. Reads work as before and carry the stale-copy warning.
+- `recommend_model` still works on Studio 3.x. It recommends from hardware alone and notes that Studio usage was not used.
+- `get_status(scope=server)` reports `studio_target` as 2.4 to 2.9.x, with Studio 3.x detected and refused.
+- README warning about Studio 3.x, and the write conditions in the README name the Studio version check and Windows and Linux.
+
+Model, service, Nexus, calibration and Bloom tools are unchanged.
+
+## v6.2.1 - 2026-11-02
+
+Glama tool-quality pass. 61 tools become 54.
+
+### Breaking changes
+
+| Old name | New name |
+|---|---|
+| `get_server_status` | `get_status(scope=server)` |
+| `analyse_msty_health` | `get_status(scope=install)` |
+| `get_service_status` | `get_status(scope=services)` |
+| Combined health call | `get_status(scope=all)` (default) |
+| `chat_with_local_model` | `chat_with_model(backend=local_ai)` |
+| `chat_with_mlx_model` | `chat_with_model(backend=mlx)` |
+| `chat_with_llamacpp_model` | `chat_with_model(backend=llamacpp)` |
+| `list_available_models` | `list_models(backend=all)` |
+| `list_mlx_models` | `list_models(backend=mlx)` |
+| `list_llamacpp_models` | `list_models(backend=llamacpp)` |
+| `list_nexus_models` | `list_models(backend=nexus)` |
+| `detect_msty_installation` | `get_msty_installation` |
+| `detect_nexus` | `get_nexus` |
+| `analyse_conversation_patterns` | `get_conversation_patterns` |
+
+`ask_frontier` is unchanged.
+
+### Added
+
+- MCP `ToolAnnotations` (`readOnlyHint`, `destructiveHint`, `idempotentHint`, `openWorldHint`, `title`) on every tool.
+- TDQS-style descriptions and a description on every parameter.
+- `glama.json` maintainer record and a `python:3.12-slim` Dockerfile for Glama Firecracker builds.
+- Persona `colorIcon` allow-list (`{"name":"tabler:[a-z0-9-]{1,64}"}`). Other values fall back to `tabler:user-scan`.
+- Length caps on persona name (100), title (100), description (500), system prompt (20000) and prompt text (20000). Over-length fields are refused with a clear error.
+- `install_model` pulls only bare Ollama names (`registry.ollama.ai`) and `hf.co/` paths. Other hosts are refused.
+
+### Changed
+
+- README tool list, architecture phase counts and Git install pin match 54 tools and v6.2.1.
+
 ## v6.2.0 - 2026-11-01
 
 Hybrid local and cloud tools on top of 6.1.5.

msty_admin_mcp-6.2.2/Dockerfile

@@ -0,0 +1,13 @@
+FROM python:3.12-slim@sha256:aa943ce411423c7582288f18707ff66eddcbaeb510502d747a72aeb34f30d159
+
+WORKDIR /app
+
+COPY pyproject.toml README.md LICENSE ./
+COPY src ./src
+
+RUN pip install --no-cache-dir . \
+    && useradd --create-home --uid 10001 app
+
+USER app
+
+CMD ["msty-admin-mcp"]

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.5
 Name: msty-admin-mcp
-Version: 6.2.0
-Summary: Comprehensive MCP server for administering Msty Studio Desktop 2.9+ with 61 tools, hardware-aware model install, Studio writes with confirm, insights analytics, Nexus bridge, Bloom evaluation, and a Cursor/Grok/Claude handoff for local models
+Version: 6.2.2
+Summary: Comprehensive MCP server for administering Msty Studio Desktop 2.9+ with 54 tools, hardware-aware model install, Studio writes with confirm, insights analytics, Nexus bridge, Bloom evaluation, and a Cursor/Grok/Claude handoff for local models
 Author-email: M-Pineapple 🍍 <128090761+M-Pineapple@users.noreply.github.com>
 License: MIT
 License-File: LICENSE
@@ -22,7 +22,7 @@ Description-Content-Type: text/markdown
 
 <img src="assets/logo.svg" alt="Msty Admin MCP logo" width="240"/>
 
-# Msty Admin MCP v6.2.0
+# Msty Admin MCP v6.2.2
 
 [![Licence: MIT](https://img.shields.io/github/license/M-Pineapple/msty-admin-mcp)](LICENSE)
 [![Python](https://img.shields.io/badge/Python-3.10+-3776AB?logo=python&logoColor=white)](https://www.python.org)
@@ -33,6 +33,9 @@ Description-Content-Type: text/markdown
 
 > Unofficial community project. Not affiliated with, or endorsed by, Msty. Msty is a trademark of its owner.
 
+> [!WARNING]
+> **Msty Studio 3.x is not supported yet.** Studio 3.0 can move workspaces to file-based storage, which this release cannot read safely. From v6.2.2 the server detects Studio 3.x and turns off every Studio read and write, so it never reports a stale copy as current or writes a row Studio 3.0 may reject. Model, service, Nexus, calibration and Bloom tools still work. Supported Studio versions are 2.4 to 2.9.x. Persona and prompt writes need a detected Studio 2.4 to 2.9.x, so they are refused when the version cannot be read. On Windows and Linux the Studio version cannot be read yet, so persona and prompt writes are refused there. Import the saved JSON pack in Persona Studio instead. If you rely on the Studio data tools, stay on the Studio 2.x update line for now. Studio 2.x installs stay on that line until you choose the 3.x track.
+
 <img src="assets/demo.gif" alt="msty-admin-mcp answering initialize and listing its tools" width="720"/>
 
 **Let the AI you already use plan, check and look after your local AI in Msty Studio, and let your local model ask for stronger help only when it needs it.**
@@ -111,7 +114,7 @@ For this to work, the chat uses a small toolset holding only the Frontier helper
 
 ## What you can do
 
-Once the server is added to your AI app, ask as you normally would. You never need to name a tool, because the AI picks the right one. The groups below cover all 61 tools. The full names are in [Tools summary](#tools-summary-61-in-total).
+Once the server is added to your AI app, ask as you normally would. You never need to name a tool, because the AI picks the right one. The groups below cover all 54 tools. The full names are in [Tools summary](#tools-summary-54-in-total).
 
 ### Know your machine and pick a model
 
@@ -204,7 +207,7 @@ Msty Studio users who want their usual AI app to help choose, test and organise
 
 - Inventory, analytics and free-form SQL stay **read only**. Credential columns always read as NULL. Secrets typed into free text, such as a chat message or a note, are out of scope.
 - Write tools always save a JSON pack under `~/.msty-admin/imports/` (mode `0600`, directory `0700`). `import_tool_config` and the tools, toolsets and skills keys of `import_studio_pack` never write the Studio database. Import that pack in Toolbox. Configs with `command`, `args` or `env` are flagged as running a program on your machine.
-- Personas and prompts insert into Studio SQLite only when both `confirm=true` and the operator env `MSTY_ALLOW_STUDIO_WRITES=1` are set. The LLM cannot set that env through a tool argument. Writes take a SQLite backup (including WAL) under `~/.msty-admin/backups/` and refuse while Msty Studio is running against the live database. Tests use fixture copies, never the live DB.
+- Personas and prompts insert into Studio SQLite only when `confirm=true`, the operator env `MSTY_ALLOW_STUDIO_WRITES=1` and a detected Studio 2.4 to 2.9.x are all present. On Windows and Linux the Studio version cannot be read yet, so these writes are refused there. The LLM cannot set that env through a tool argument. Writes take a SQLite backup (including WAL) under `~/.msty-admin/backups/` and refuse while Msty Studio is running against the live database. Tests use fixture copies, never the live DB.
 - `query_local_ai_service` is allow-listed (tags, version, show, chat). Pull and create go through `install_model` and `set_model_settings`, which also need `confirm=true`.
 - Calibration and Bloom history live under `~/.msty-admin/`, not in Studio.
 - Frontier needs `CURSOR_API_KEY` for Cursor, `XAI_API_KEY` for Grok, or `ANTHROPIC_API_KEY` for Claude. Keys stay in Msty's toolbox environment and never belong in this repository. In ask mode the Cursor agent does not send mail, move or delete files, or use trading tools.
@@ -235,7 +238,7 @@ If you already have Python 3.10+ as `python3`, `python3 -m venv .venv` is enough
 
 If `pip install msty-admin-mcp` cannot find the package, install the release straight from GitHub:
 ```bash
-pip install "msty-admin-mcp @ git+https://github.com/M-Pineapple/msty-admin-mcp@v6.2.0"
+pip install "msty-admin-mcp @ git+https://github.com/M-Pineapple/msty-admin-mcp@v6.2.2"
 ```
 
 ### With HTTP Transport
@@ -341,14 +344,14 @@ Msty Studio Desktop
          ↓
 
 MCP Server (stdio / HTTP)
-├── Phase 1: Foundational (6)
+├── Phase 1: Foundational (5)
 ├── Phase 2: Configuration (5)
-├── Phase 3: Service Integration (14)
+├── Phase 3: Service Integration (9)
 ├── Phase 4: Intelligence (5)
 ├── Phase 5: Calibration (5)
 ├── Phase 6: Bloom Evaluation (6)
 ├── Phase 7: Studio Inventory (15)
-├── Phase 8: Nexus Bridge (4)
+├── Phase 8: Nexus Bridge (3)
 └── Phase 9: Frontier handoff (1)
 ```
 
@@ -360,7 +363,7 @@ A local model does the ordinary work. When a question is past it, that model cal
 
 Msty Studio has no hook API. `escalate_if_needed` and the "Local with frontier escalation" persona are the honest equivalent: the local model must emit a real tool call.
 
-Use a **separate one-tool server** for the local-model chat. The full admin server has 61 tools. A small model will call those other tools (GitHub, search, schema, and so on) instead of waiting for the answer.
+Use a **separate one-tool server** for the local-model chat. The full admin server has 54 tools. A small model will call those other tools (GitHub, search, schema, and so on) instead of waiting for the answer.
 
 ### Fresh install
 
@@ -411,35 +414,34 @@ The admin server still has the Studio inventory tools. Add the Frontier server a
 
 ---
 
-## Tools summary (61 in total)
+## Tools summary (54 in total)
 
-### Phase 1: Foundational (6)
-- `detect_msty_installation`: app, version, DB path, ports
+### Phase 1: Foundational (5)
+- `get_msty_installation`: app, version, DB path, ports
 - `read_msty_database`: `stats` / `tables` / read-only SQL
   - In read-only SQL, credential columns (key, token, secret, password, auth and header names, URL, config/JSON and command-line columns) always read as NULL. Secrets typed into arbitrary free-text columns, such as a chat message or a note, are out of scope for this tool.
 - `list_configured_tools`: Toolbox MCP tools
 - `get_model_providers`: services and Studio providers (keys redacted)
-- `analyse_msty_health`: install, DB and services
-- `get_server_status`: MCP status
+- `get_status`: MCP, install and local service health. `scope` is `all`, `server`, `install` or `services`
 
 ### Phase 2: Configuration (5)
 - `export_tool_config`: export a tool by name/id
 - `sync_claude_preferences`: Claude to Msty delta report
-- `generate_persona`: persona JSON pack. SQLite insert needs `confirm=true` and `MSTY_ALLOW_STUDIO_WRITES=1`
+- `generate_persona`: persona JSON pack. SQLite insert needs `confirm=true`, `MSTY_ALLOW_STUDIO_WRITES=1` and a detected Studio 2.4 to 2.9.x, so it is refused on Windows and Linux
 - `import_tool_config`: JSON pack for Toolbox. Never writes the Studio database
 - `import_studio_pack`: tools, toolsets and skills stay in the JSON pack. Personas and prompts need the env gate
 
-### Phase 3: Service Integration (14)
-- `get_service_status`, `list_available_models`, `query_local_ai_service` (allow-listed)
-- `chat_with_local_model`, `get_hardware_profile`, `recommend_model`
+### Phase 3: Service Integration (9)
+- `list_models`: Local AI, MLX, llama.cpp, Vibe and Nexus. `backend` defaults to `all`
+- `query_local_ai_service` (allow-listed)
+- `chat_with_model`: Local AI, MLX or llama.cpp. `backend` defaults to `local_ai`
+- `get_hardware_profile`, `recommend_model`
 - `install_model`, `set_model_settings`
-- `list_mlx_models`, `chat_with_mlx_model`
-- `list_llamacpp_models`, `chat_with_llamacpp_model`
 - `get_vibe_proxy_status`, `query_vibe_proxy`
 
 ### Phase 4: Intelligence (5)
 - `get_model_performance_metrics`: from Studio `insightsLanguageModelUsage`
-- `analyse_conversation_patterns`
+- `get_conversation_patterns`
 - `compare_model_responses`
 - `optimise_knowledge_stacks`: live Next Gen / draft / compose status
 - `suggest_persona_improvements`
@@ -460,8 +462,8 @@ The admin server still has the Studio inventory tools. Add the Frontier server a
 - `list_context_studio`, `list_agent_plans`
 - `get_studio_entity` (plural keys such as `personas`, plus singular aliases such as `persona`), `get_schema_migrations`, `export_workflow_pack`
 
-### Phase 8: Nexus Bridge (4)
-- `detect_nexus`, `list_nexus_models`, `query_nexus`, `get_insights_usage`
+### Phase 8: Nexus Bridge (3)
+- `get_nexus`, `query_nexus`, `get_insights_usage`
 
 ### Phase 9: Frontier handoff (1)
 - `ask_frontier`: hand one question to Cursor (default), Grok, or Claude and return the answer in the Msty chat. Use the one-tool server (`msty-frontier` or `python -m src.frontier_main`) for a local-model chat. See [Hybrid local model and Cursor agent](#hybrid-local-model-and-cursor-agent).
@@ -498,10 +500,10 @@ v6 ships unit tests for path resolution, read-only SQL guards, redaction, invent
 **A**: No. v6 targets Studio 2.9+ (works on 2.4+ layouts when the File System DB is present). Use v4.x for older process-based detection.
 
 ### Q: Does this write to the Studio database?
-**A**: Inventory and analytics stay read-only. Write tools always save a JSON pack under `~/.msty-admin/imports/`. Tools, toolsets and skills are import-file only. Personas and prompts insert into Studio SQLite only with `confirm=true` and `MSTY_ALLOW_STUDIO_WRITES=1` in the MCP server environment, after a timestamped SQLite backup that includes WAL, and they refuse while Msty Studio is running against the live database. Quit Studio first, or import the JSON pack in Persona Studio or Toolbox.
+**A**: Inventory and analytics stay read-only. Write tools always save a JSON pack under `~/.msty-admin/imports/`. Tools, toolsets and skills are import-file only. Personas and prompts insert into Studio SQLite only with `confirm=true`, `MSTY_ALLOW_STUDIO_WRITES=1` in the MCP server environment and a detected Studio 2.4 to 2.9.x, after a timestamped SQLite backup that includes WAL, and they refuse while Msty Studio is running against the live database. Quit Studio first, or import the JSON pack in Persona Studio or Toolbox. On Windows and Linux the Studio version cannot be read yet, so these writes are refused there.
 
 ### Q: What is the Nexus integration?
-**A**: [Msty Nexus](https://msty.ai/nexus/) is the local runtime/gateway. Phase 8 detects it and lists models via its OpenAI-compatible endpoint (`MSTY_NEXUS_PORT`).
+**A**: [Msty Nexus](https://msty.ai/nexus/) is the local runtime/gateway. `get_nexus` detects it. `list_models` with `backend=nexus` lists models on its OpenAI-compatible endpoint (`MSTY_NEXUS_PORT`).
 
 ---
 

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/README.md

@@ -2,7 +2,7 @@
 
 <img src="assets/logo.svg" alt="Msty Admin MCP logo" width="240"/>
 
-# Msty Admin MCP v6.2.0
+# Msty Admin MCP v6.2.2
 
 [![Licence: MIT](https://img.shields.io/github/license/M-Pineapple/msty-admin-mcp)](LICENSE)
 [![Python](https://img.shields.io/badge/Python-3.10+-3776AB?logo=python&logoColor=white)](https://www.python.org)
@@ -13,6 +13,9 @@
 
 > Unofficial community project. Not affiliated with, or endorsed by, Msty. Msty is a trademark of its owner.
 
+> [!WARNING]
+> **Msty Studio 3.x is not supported yet.** Studio 3.0 can move workspaces to file-based storage, which this release cannot read safely. From v6.2.2 the server detects Studio 3.x and turns off every Studio read and write, so it never reports a stale copy as current or writes a row Studio 3.0 may reject. Model, service, Nexus, calibration and Bloom tools still work. Supported Studio versions are 2.4 to 2.9.x. Persona and prompt writes need a detected Studio 2.4 to 2.9.x, so they are refused when the version cannot be read. On Windows and Linux the Studio version cannot be read yet, so persona and prompt writes are refused there. Import the saved JSON pack in Persona Studio instead. If you rely on the Studio data tools, stay on the Studio 2.x update line for now. Studio 2.x installs stay on that line until you choose the 3.x track.
+
 <img src="assets/demo.gif" alt="msty-admin-mcp answering initialize and listing its tools" width="720"/>
 
 **Let the AI you already use plan, check and look after your local AI in Msty Studio, and let your local model ask for stronger help only when it needs it.**
@@ -91,7 +94,7 @@ For this to work, the chat uses a small toolset holding only the Frontier helper
 
 ## What you can do
 
-Once the server is added to your AI app, ask as you normally would. You never need to name a tool, because the AI picks the right one. The groups below cover all 61 tools. The full names are in [Tools summary](#tools-summary-61-in-total).
+Once the server is added to your AI app, ask as you normally would. You never need to name a tool, because the AI picks the right one. The groups below cover all 54 tools. The full names are in [Tools summary](#tools-summary-54-in-total).
 
 ### Know your machine and pick a model
 
@@ -184,7 +187,7 @@ Msty Studio users who want their usual AI app to help choose, test and organise
 
 - Inventory, analytics and free-form SQL stay **read only**. Credential columns always read as NULL. Secrets typed into free text, such as a chat message or a note, are out of scope.
 - Write tools always save a JSON pack under `~/.msty-admin/imports/` (mode `0600`, directory `0700`). `import_tool_config` and the tools, toolsets and skills keys of `import_studio_pack` never write the Studio database. Import that pack in Toolbox. Configs with `command`, `args` or `env` are flagged as running a program on your machine.
-- Personas and prompts insert into Studio SQLite only when both `confirm=true` and the operator env `MSTY_ALLOW_STUDIO_WRITES=1` are set. The LLM cannot set that env through a tool argument. Writes take a SQLite backup (including WAL) under `~/.msty-admin/backups/` and refuse while Msty Studio is running against the live database. Tests use fixture copies, never the live DB.
+- Personas and prompts insert into Studio SQLite only when `confirm=true`, the operator env `MSTY_ALLOW_STUDIO_WRITES=1` and a detected Studio 2.4 to 2.9.x are all present. On Windows and Linux the Studio version cannot be read yet, so these writes are refused there. The LLM cannot set that env through a tool argument. Writes take a SQLite backup (including WAL) under `~/.msty-admin/backups/` and refuse while Msty Studio is running against the live database. Tests use fixture copies, never the live DB.
 - `query_local_ai_service` is allow-listed (tags, version, show, chat). Pull and create go through `install_model` and `set_model_settings`, which also need `confirm=true`.
 - Calibration and Bloom history live under `~/.msty-admin/`, not in Studio.
 - Frontier needs `CURSOR_API_KEY` for Cursor, `XAI_API_KEY` for Grok, or `ANTHROPIC_API_KEY` for Claude. Keys stay in Msty's toolbox environment and never belong in this repository. In ask mode the Cursor agent does not send mail, move or delete files, or use trading tools.
@@ -215,7 +218,7 @@ If you already have Python 3.10+ as `python3`, `python3 -m venv .venv` is enough
 
 If `pip install msty-admin-mcp` cannot find the package, install the release straight from GitHub:
 ```bash
-pip install "msty-admin-mcp @ git+https://github.com/M-Pineapple/msty-admin-mcp@v6.2.0"
+pip install "msty-admin-mcp @ git+https://github.com/M-Pineapple/msty-admin-mcp@v6.2.2"
 ```
 
 ### With HTTP Transport
@@ -321,14 +324,14 @@ Msty Studio Desktop
          ↓
 
 MCP Server (stdio / HTTP)
-├── Phase 1: Foundational (6)
+├── Phase 1: Foundational (5)
 ├── Phase 2: Configuration (5)
-├── Phase 3: Service Integration (14)
+├── Phase 3: Service Integration (9)
 ├── Phase 4: Intelligence (5)
 ├── Phase 5: Calibration (5)
 ├── Phase 6: Bloom Evaluation (6)
 ├── Phase 7: Studio Inventory (15)
-├── Phase 8: Nexus Bridge (4)
+├── Phase 8: Nexus Bridge (3)
 └── Phase 9: Frontier handoff (1)
 ```
 
@@ -340,7 +343,7 @@ A local model does the ordinary work. When a question is past it, that model cal
 
 Msty Studio has no hook API. `escalate_if_needed` and the "Local with frontier escalation" persona are the honest equivalent: the local model must emit a real tool call.
 
-Use a **separate one-tool server** for the local-model chat. The full admin server has 61 tools. A small model will call those other tools (GitHub, search, schema, and so on) instead of waiting for the answer.
+Use a **separate one-tool server** for the local-model chat. The full admin server has 54 tools. A small model will call those other tools (GitHub, search, schema, and so on) instead of waiting for the answer.
 
 ### Fresh install
 
@@ -391,35 +394,34 @@ The admin server still has the Studio inventory tools. Add the Frontier server a
 
 ---
 
-## Tools summary (61 in total)
+## Tools summary (54 in total)
 
-### Phase 1: Foundational (6)
-- `detect_msty_installation`: app, version, DB path, ports
+### Phase 1: Foundational (5)
+- `get_msty_installation`: app, version, DB path, ports
 - `read_msty_database`: `stats` / `tables` / read-only SQL
   - In read-only SQL, credential columns (key, token, secret, password, auth and header names, URL, config/JSON and command-line columns) always read as NULL. Secrets typed into arbitrary free-text columns, such as a chat message or a note, are out of scope for this tool.
 - `list_configured_tools`: Toolbox MCP tools
 - `get_model_providers`: services and Studio providers (keys redacted)
-- `analyse_msty_health`: install, DB and services
-- `get_server_status`: MCP status
+- `get_status`: MCP, install and local service health. `scope` is `all`, `server`, `install` or `services`
 
 ### Phase 2: Configuration (5)
 - `export_tool_config`: export a tool by name/id
 - `sync_claude_preferences`: Claude to Msty delta report
-- `generate_persona`: persona JSON pack. SQLite insert needs `confirm=true` and `MSTY_ALLOW_STUDIO_WRITES=1`
+- `generate_persona`: persona JSON pack. SQLite insert needs `confirm=true`, `MSTY_ALLOW_STUDIO_WRITES=1` and a detected Studio 2.4 to 2.9.x, so it is refused on Windows and Linux
 - `import_tool_config`: JSON pack for Toolbox. Never writes the Studio database
 - `import_studio_pack`: tools, toolsets and skills stay in the JSON pack. Personas and prompts need the env gate
 
-### Phase 3: Service Integration (14)
-- `get_service_status`, `list_available_models`, `query_local_ai_service` (allow-listed)
-- `chat_with_local_model`, `get_hardware_profile`, `recommend_model`
+### Phase 3: Service Integration (9)
+- `list_models`: Local AI, MLX, llama.cpp, Vibe and Nexus. `backend` defaults to `all`
+- `query_local_ai_service` (allow-listed)
+- `chat_with_model`: Local AI, MLX or llama.cpp. `backend` defaults to `local_ai`
+- `get_hardware_profile`, `recommend_model`
 - `install_model`, `set_model_settings`
-- `list_mlx_models`, `chat_with_mlx_model`
-- `list_llamacpp_models`, `chat_with_llamacpp_model`
 - `get_vibe_proxy_status`, `query_vibe_proxy`
 
 ### Phase 4: Intelligence (5)
 - `get_model_performance_metrics`: from Studio `insightsLanguageModelUsage`
-- `analyse_conversation_patterns`
+- `get_conversation_patterns`
 - `compare_model_responses`
 - `optimise_knowledge_stacks`: live Next Gen / draft / compose status
 - `suggest_persona_improvements`
@@ -440,8 +442,8 @@ The admin server still has the Studio inventory tools. Add the Frontier server a
 - `list_context_studio`, `list_agent_plans`
 - `get_studio_entity` (plural keys such as `personas`, plus singular aliases such as `persona`), `get_schema_migrations`, `export_workflow_pack`
 
-### Phase 8: Nexus Bridge (4)
-- `detect_nexus`, `list_nexus_models`, `query_nexus`, `get_insights_usage`
+### Phase 8: Nexus Bridge (3)
+- `get_nexus`, `query_nexus`, `get_insights_usage`
 
 ### Phase 9: Frontier handoff (1)
 - `ask_frontier`: hand one question to Cursor (default), Grok, or Claude and return the answer in the Msty chat. Use the one-tool server (`msty-frontier` or `python -m src.frontier_main`) for a local-model chat. See [Hybrid local model and Cursor agent](#hybrid-local-model-and-cursor-agent).
@@ -478,10 +480,10 @@ v6 ships unit tests for path resolution, read-only SQL guards, redaction, invent
 **A**: No. v6 targets Studio 2.9+ (works on 2.4+ layouts when the File System DB is present). Use v4.x for older process-based detection.
 
 ### Q: Does this write to the Studio database?
-**A**: Inventory and analytics stay read-only. Write tools always save a JSON pack under `~/.msty-admin/imports/`. Tools, toolsets and skills are import-file only. Personas and prompts insert into Studio SQLite only with `confirm=true` and `MSTY_ALLOW_STUDIO_WRITES=1` in the MCP server environment, after a timestamped SQLite backup that includes WAL, and they refuse while Msty Studio is running against the live database. Quit Studio first, or import the JSON pack in Persona Studio or Toolbox.
+**A**: Inventory and analytics stay read-only. Write tools always save a JSON pack under `~/.msty-admin/imports/`. Tools, toolsets and skills are import-file only. Personas and prompts insert into Studio SQLite only with `confirm=true`, `MSTY_ALLOW_STUDIO_WRITES=1` in the MCP server environment and a detected Studio 2.4 to 2.9.x, after a timestamped SQLite backup that includes WAL, and they refuse while Msty Studio is running against the live database. Quit Studio first, or import the JSON pack in Persona Studio or Toolbox. On Windows and Linux the Studio version cannot be read yet, so these writes are refused there.
 
 ### Q: What is the Nexus integration?
-**A**: [Msty Nexus](https://msty.ai/nexus/) is the local runtime/gateway. Phase 8 detects it and lists models via its OpenAI-compatible endpoint (`MSTY_NEXUS_PORT`).
+**A**: [Msty Nexus](https://msty.ai/nexus/) is the local runtime/gateway. `get_nexus` detects it. `list_models` with `backend=nexus` lists models on its OpenAI-compatible endpoint (`MSTY_NEXUS_PORT`).
 
 ---
 

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/docs/BLOOM_GUIDE.md

@@ -484,7 +484,7 @@ Or add it to your Claude Desktop MCP configuration:
 **Fix**: Run a quick chat with the model first to load it into memory:
 
 ```
-chat_with_local_model(model="qwen2.5:7b", messages=[{"role": "user", "content": "hello"}])
+chat_with_model(model="qwen2.5:7b", messages=[{"role": "user", "content": "hello"}], backend="local_ai")
 ```
 
 Then run the evaluation.

msty_admin_mcp-6.2.2/glama.json

@@ -0,0 +1,4 @@
+{
+  "$schema": "https://glama.ai/mcp/schemas/server.json",
+  "maintainers": ["M-Pineapple"]
+}

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/pyproject.toml

@@ -4,8 +4,8 @@ build-backend = "hatchling.build"
 
 [project]
 name = "msty-admin-mcp"
-version = "6.2.0"
-description = "Comprehensive MCP server for administering Msty Studio Desktop 2.9+ with 61 tools, hardware-aware model install, Studio writes with confirm, insights analytics, Nexus bridge, Bloom evaluation, and a Cursor/Grok/Claude handoff for local models"
+version = "6.2.2"
+description = "Comprehensive MCP server for administering Msty Studio Desktop 2.9+ with 54 tools, hardware-aware model install, Studio writes with confirm, insights analytics, Nexus bridge, Bloom evaluation, and a Cursor/Grok/Claude handoff for local models"
 readme = "README.md"
 requires-python = ">=3.10"
 license = {text = "MIT"}

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/src/db.py

@@ -13,7 +13,14 @@ import time
 from pathlib import Path
 from typing import Any, Dict, List, Optional, Sequence, Tuple
 
-from src.paths import detect_msty_installation
+from src.paths import (
+    LAYOUT_UNKNOWN,
+    LAYOUT_UNSUPPORTED,
+    UnsupportedStudioLayout,
+    detect_msty_installation,
+    unknown_version_read_warning,
+    unsupported_layout_message,
+)
 
 # Only allow SELECT / WITH / PRAGMA for arbitrary query tool
 _SAFE_SQL = re.compile(r"^\s*(SELECT|WITH|PRAGMA)\b", re.IGNORECASE | re.DOTALL)
@@ -393,11 +400,32 @@ def names_credential_column(query: str) -> bool:
     return False
 
 
+def with_read_warning(payload: Any) -> Any:
+    """Add a stale-copy warning to a read result when the Studio version is unknown."""
+    if not isinstance(payload, dict):
+        return payload
+    try:
+        status = detect_msty_installation().layout_status
+    except Exception:
+        status = LAYOUT_UNKNOWN
+    if status == LAYOUT_UNKNOWN:
+        payload = dict(payload)
+        payload["warning"] = unknown_version_read_warning()
+    return payload
+
+
 def get_database_path(explicit: Optional[str] = None) -> Optional[Path]:
+    """Resolve the Studio SQLite path for reads.
+
+    Raises UnsupportedStudioLayout on Studio 3.x, even for an explicit path,
+    so no Studio-storage read runs against a layout this release cannot read.
+    """
+    install = detect_msty_installation()
+    if install.layout_status == LAYOUT_UNSUPPORTED:
+        raise UnsupportedStudioLayout(unsupported_layout_message(install.version))
     if explicit:
         path = Path(explicit)
         return path if path.exists() else None
-    install = detect_msty_installation()
     if install.database_path:
         return Path(install.database_path)
     return None

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/src/install.py

@@ -70,6 +70,39 @@ def leftover_scheme(name: str) -> bool:
     return "://" in _strip_uri_schemes((name or "").strip().lower())
 
 
+def validate_pull_source(name: str) -> Optional[str]:
+    """Refuse pull names that point at a host other than Ollama or hf.co/.
+
+    Bare names (llama3.2:3b) and registry.ollama.ai/... are allowed. Hugging
+    Face paths must start with hf.co/. Any other dotted first path segment is
+    treated as a host and refused.
+    """
+    text = _strip_uri_schemes((name or "").strip().lower())
+    if not text:
+        return "model is empty"
+    if text.startswith("hf.co/"):
+        rest = text[6:]
+        if not rest or rest.startswith("/") or ".." in rest:
+            return "hf.co/ model path is empty or unsafe"
+        return None
+    if _REGISTRY_HOST_RE.match(text):
+        return None
+    if text.count("/") >= 2:
+        return (
+            "install_model only pulls bare Ollama names (registry.ollama.ai) "
+            "or hf.co/ paths. Names with three or more parts must start with "
+            "registry.ollama.ai/ or hf.co/."
+        )
+    if "/" in text:
+        host = text.split("/", 1)[0]
+        if "." in host or ":" in host or "[" in host or host == "localhost":
+            return (
+                "install_model only pulls bare Ollama names (registry.ollama.ai) "
+                f"or hf.co/ paths. Refused host {host!r}."
+            )
+    return None
+
+
 def normalise_model_name(name: str) -> str:
     """Lowercase, strip scheme and registry/library prefixes, add :latest when untagged."""
     text = _strip_uri_schemes((name or "").strip().lower())
@@ -204,6 +237,9 @@ def pull_model(
     name_error = validate_model_name(cleaned, "model")
     if name_error:
         return {"success": False, "error": name_error, "wrote": False}
+    source_error = validate_pull_source(cleaned)
+    if source_error:
+        return {"success": False, "error": source_error, "wrote": False}
     if service.strip().lower() in {"mlx", "llamacpp", "llama.cpp"}:
         return {
             "success": False,
@@ -329,8 +365,11 @@ def create_variant(
     """POST /api/create with structured fields. Modelfile text is preview only."""
     base = (base_model or "").strip()
     variant = (variant_name or "").strip()
-    name_error = validate_model_name(base, "base_model") or validate_model_name(
-        variant, "variant_name"
+    name_error = (
+        validate_model_name(base, "base_model")
+        or validate_model_name(variant, "variant_name")
+        or validate_pull_source(base)
+        or validate_pull_source(variant)
     )
     if name_error:
         return {"success": False, "error": name_error, "wrote": False}

{msty_admin_mcp-6.2.0 → msty_admin_mcp-6.2.2}/src/intelligence/__init__.py

@@ -3,7 +3,7 @@ Phase 4: Intelligence Layer / Analytics Tools for Msty Admin MCP
 
 Tools:
 - get_model_performance_metrics
-- analyse_conversation_patterns
+- get_conversation_patterns
 - compare_model_responses
 - optimise_knowledge_stacks
 - suggest_persona_improvements