PyPI - dynamitecircle - Versions diffs - 2.0.1__tar.gz → 2.0.4__tar.gz - Mend

dynamitecircle 2.0.1tar.gz → 2.0.4tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (15) hide show

{dynamitecircle-2.0.1 → dynamitecircle-2.0.4}/CHANGELOG.md RENAMED Viewed

@@ -14,6 +14,56 @@ the public Python API surface (`dc.DC`, `dc.DCError`, `dc.Result`,
 ---
+## [Unreleased]
+## [2.0.4] – 2026-06-16
+### Added
+- **MCP tool annotations** — every tool now advertises `readOnlyHint` /
+  `destructiveHint` / `openWorldHint`, so clients can auto-approve reads and
+  flag writes. The read/write split is derived from each command's actual HTTP
+  verb and guarded by a test that re-derives it from source (no silent drift).
+  Write tools also carry a "⚠️ Write operation" note in their description,
+  mirroring the hosted MCP server.
+- **Structured tool output** — object results (the list envelope
+  `{items, count, cursor, has_more}` and single records) are also returned as
+  MCP `structuredContent`, so structure-aware clients don't have to re-parse
+  the text. No `outputSchema` is declared (it would strictly validate every
+  response shape); lists/scalars remain text-only.
+### Changed
+- **BREAKING (MCP interface) — tool names and fields are now snake_case.**
+  MCP tool names switched from kebab-case to snake_case
+  (`mcp__dc__trip-create` → `mcp__dc__trip_create`), and tool argument fields
+  likewise (`event-id` → `event_id`, `start-date` → `start_date`). This aligns
+  the agent-facing MCP surface with the snake_case Python library
+  (`dc.trip_create()`) and the prevailing MCP tool-naming convention. The
+  **CLI is unchanged** (`dc trip-create`, `--start-date`). Wildcard allow-rules
+  like `mcp__dc__*` are unaffected; any config pinning a specific kebab tool
+  name must update it. `_call_tool` still accepts the legacy kebab form as a
+  transitional fallback.
+### Docs / packaging
+- Document the hosted MCP (`https://api.dynamitecircle.com/mcp`) and an
+  auto-updating `uvx` PyPI config; fix the `.venv/bin/python3` configs to bare
+  `python3`; unify `py/SKILL.md` (local client + hosted MCP) and add
+  `DC/SKILL.md`; add `tests/test_version_sync.py` guarding manifest/config
+  drift from `DC_API_VERSION`.
+- **One-click install** — "Add to Cursor" and "Install in VS Code" badges for
+  the hosted MCP, a native VS Code section (`.vscode/mcp.json` + `code --add-mcp`),
+  and a `workflows`-first note for the ~80-tool surface.
+- **`.mcpb` bundle** — the release workflow now builds a `dc.mcpb` (one-click
+  Claude Desktop install) via `mcpb pack` and attaches it to the GitHub release;
+  `.mcpbignore` keeps the bundle lean.
+- **Official MCP Registry** — add `server.json` (`io.github.dynamitecircle/dc`,
+  listing the hosted streamable-HTTP endpoint) and a separate, isolated
+  `publish-mcp.yml` workflow that publishes it via GitHub OIDC on `v*` tags.
+  The registry feeds VS Code, Smithery, PulseMCP, and others. `server.json`
+  version is guarded by `tests/test_version_sync.py`.
 ## [2.0.1] – 2026-06-16
 ### Changed
@@ -1071,7 +1121,8 @@ Tag exists but no PyPI release. Replaced by 1.6.3.
 ---
-[Unreleased]: https://github.com/dynamitecircle/dc/compare/v2.0.1...HEAD
+[Unreleased]: https://github.com/dynamitecircle/dc/compare/v2.0.4...HEAD
+[2.0.4]: https://github.com/dynamitecircle/dc/compare/v2.0.1...v2.0.4
 [2.0.1]: https://github.com/dynamitecircle/dc/releases/tag/v2.0.1
 [1.23.3]: https://github.com/dynamitecircle/dc/releases/tag/v1.23.3
 [1.22.6]: https://github.com/dynamitecircle/dc/releases/tag/v1.22.6

{dynamitecircle-2.0.1 → dynamitecircle-2.0.4}/PKG-INFO RENAMED Viewed

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dynamitecircle
-Version: 2.0.1
+Version: 2.0.4
 Summary: Official Python client for the public Dynamite Circle Member API. CLI, library, Agent Skill, and MCP server in one file.
 Project-URL: Homepage, https://github.com/dynamitecircle/dc
 Project-URL: Repository, https://github.com/dynamitecircle/dc
@@ -153,6 +153,37 @@ The same `dc.py` file is shipped as **four** integrations — pick whichever fit
 The `mcp` package is **lazy-imported** — Agent Skill / CLI / Python-library users never need it.
+### Don't want to install anything? Use the hosted MCP
+The DC Member API is **also a hosted MCP server** — a remote endpoint you point your AI app at, with nothing to clone, install, or update:
+```
+https://api.dynamitecircle.com/mcp        (Streamable HTTP)
+```
+[![Add to Cursor](https://cursor.com/deeplink/mcp-install-dark.svg)](cursor://anysphere.cursor-deeplink/mcp/install?name=dc&config=eyJ1cmwiOiJodHRwczovL2FwaS5keW5hbWl0ZWNpcmNsZS5jb20vbWNwIn0=)
+&nbsp;
+[![Install in VS Code](https://img.shields.io/badge/VS_Code-Install_DC-007ACC?logo=visualstudiocode&logoColor=white)](https://insiders.vscode.dev/redirect/mcp/install?name=dc&config=%7B%22url%22%3A%22https%3A%2F%2Fapi.dynamitecircle.com%2Fmcp%22%7D)
+- **Claude Code:** `claude mcp add --transport http dc https://api.dynamitecircle.com/mcp`
+- **Claude (web / Desktop / mobile):** Settings → Connectors → **Add custom connector** → paste the URL → **Connect**
+- **VS Code:** click the badge above, or `code --add-mcp '{"name":"dc","url":"https://api.dynamitecircle.com/mcp"}'`
+- **Cursor:** click the badge above, or add a custom MCP connector pointing at the URL
+- **ChatGPT / any other MCP app:** add a custom / remote MCP connector pointing at the URL above
+When prompted, **sign in with your DC account** (one-click OAuth) — or send a `dk_` key as `Authorization: Bearer`. The server is always on the current API version, so there's nothing to keep up to date.
+**Hosted MCP vs. this client — which do I want?**
+| You want… | Use |
+|---|---|
+| Zero install, always-current, just MCP tools in a chat app | **Hosted MCP** (URL above) |
+| To call DC from your own Python code (`from dynamitecircle import DC`) | **This client** (library) |
+| A local CLI (`dc profile`) or shell scripting | **This client** (CLI) |
+| A local stdio MCP server (offline-capable, pin a version, dev against localhost) | **This client** (`dc --mcp`) |
+The rest of this README is about **this client**. For the hosted MCP, the URL above is all you need.
 ## Features
 - **Full Member API coverage** — read + write across every public endpoint (run `python3 py/dc.py help` for the live list)
@@ -214,7 +245,18 @@ Run `python3 py/dc.py help` for the full command list.
 ## Setup per AI tool
-### Claude Code
+### Hosted MCP (no install, always current)
+Nothing to clone or update — point your tool at the remote endpoint:
+```bash
+# Claude Code
+claude mcp add --transport http dc https://api.dynamitecircle.com/mcp
+```
+For Claude web / Desktop, Cursor, ChatGPT, and other MCP apps, add a **custom / remote connector** pointing at `https://api.dynamitecircle.com/mcp` (Streamable HTTP). Sign in with your DC account (OAuth) or send a `dk_` key as a Bearer header. Discovery is published at [`/.well-known/mcp.json`](https://api.dynamitecircle.com/.well-known/mcp.json). This is the lowest-effort path; everything below runs the client **locally** instead.
+### Claude Code (local client)
 `.mcp.json` is already shipped with this repo. Open the repo with `claude`:
@@ -231,9 +273,30 @@ pip install -r py/requirements.txt
 Skill discovery (CLI + import) works automatically via `py/SKILL.md`.
+#### Auto-updating local MCP via `uvx` (PyPI)
+Prefer a local stdio server that **pulls the latest published client on every launch** with no clone to maintain? Use [`uv`](https://docs.astral.sh/uv/)'s `uvx` and pass your key in the `env` block:
+```json
+{
+  "mcpServers": {
+    "dc": {
+      "type": "stdio",
+      "command": "uvx",
+      "args": ["--refresh", "--from", "dynamitecircle[mcp]", "dc", "--mcp"],
+      "env": { "DC_API_KEY": "dk_<api-key>" }
+    }
+  }
+}
+```
+Notes: the console script is `dc` (not `dynamitecircle`), so `--from dynamitecircle[mcp]` is required — it both names the package and pulls the optional `mcp` extra. `--refresh` is what makes it auto-update; drop it to let `uvx` cache. The `DC_API_KEY` in `env` wins over any `.env.dc` (an ephemeral `uvx` install has none).
 ### Claude Desktop
-Edit your config file:
+**Easiest — one-click `.mcpb`:** download `dc.mcpb` from the [latest GitHub release](https://github.com/dynamitecircle/dc/releases/latest) and double-click it. Claude Desktop installs the `dc` server and prompts for your `DC_API_KEY` (no JSON editing). Requires Python 3.9+ with the `mcp` package available on your machine (`pip install 'dynamitecircle[mcp]'`). *(Or skip local install entirely and use the [hosted MCP connector](#hosted-mcp-no-install-always-current) above.)*
+**Manual** — edit your config file:
 - macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
 - Linux: `~/.config/Claude/claude_desktop_config.json`
@@ -330,9 +393,19 @@ Non-paginated extras (e.g. `totalUnread` on `inbox`) are passed through under an
 ## Use as a Python library
+Installed from PyPI (`pip install dynamitecircle`) — import by the package name:
+```python
+from dynamitecircle import DC
+dc = DC()
+```
+Working from a clone instead? Point `sys.path` at the `py/` folder and import as `dc`:
 ```python
 import sys
-sys.path.insert(0, "dc")
+sys.path.insert(0, "py")
 from dc import DC
 dc = DC()
@@ -366,8 +439,21 @@ The full live reference for the DC Member API — every endpoint, parameter, and
 The DC Member API ships new endpoints and refinements regularly. This skill is the official client and we update it whenever the API changes. **Plan for updates** — the skill will warn you on stderr the first time a request returns a server `X-API-Version` newer than `DC_API_VERSION`, and major-version bumps may break older clients.
+**The no-maintenance option:** if you use the [hosted MCP](#hosted-mcp-no-install-always-current) (`https://api.dynamitecircle.com/mcp`), there's nothing to update — the server always runs the current API version. The strategies below apply only to the **local client**.
 Pick whichever integration style fits your project. From simplest to most isolated:
+### 0. PyPI — `pip` / `uvx`
+```bash
+pip install dynamitecircle              # CLI + library
+pip install 'dynamitecircle[mcp]'       # + local MCP server (dc --mcp)
+```
+Then `dc setup --api-key dk_<api-key>` and `dc self-test`. Update with `pip install --upgrade dynamitecircle`. For a local MCP that auto-updates on every launch, use the [`uvx` config above](#auto-updating-local-mcp-via-uvx-pypi).
+Best for: the quickest local install, and the auto-updating `uvx` MCP setup.
 ### 1. Plain git clone — quick local use
 ```bash
@@ -506,6 +592,9 @@ dc/
 │
 │   (Future: go/, node/, rs/ folders for sister clients in other languages)
 │
+├── DC/
+│   └── SKILL.md     → ../py/SKILL.md  # human-friendly skill path (symlink)
+│
 ├── docs/                              # ← REAL design docs (canonical)
 │   ├── skill-info.md                # design rules / architecture
 │   └── mcp-info.md                    # MCP setup for every supported client

{dynamitecircle-2.0.1 → dynamitecircle-2.0.4}/README.md RENAMED Viewed

@@ -100,6 +100,37 @@ The same `dc.py` file is shipped as **four** integrations — pick whichever fit
 The `mcp` package is **lazy-imported** — Agent Skill / CLI / Python-library users never need it.
+### Don't want to install anything? Use the hosted MCP
+The DC Member API is **also a hosted MCP server** — a remote endpoint you point your AI app at, with nothing to clone, install, or update:
+```
+https://api.dynamitecircle.com/mcp        (Streamable HTTP)
+```
+[![Add to Cursor](https://cursor.com/deeplink/mcp-install-dark.svg)](cursor://anysphere.cursor-deeplink/mcp/install?name=dc&config=eyJ1cmwiOiJodHRwczovL2FwaS5keW5hbWl0ZWNpcmNsZS5jb20vbWNwIn0=)
+&nbsp;
+[![Install in VS Code](https://img.shields.io/badge/VS_Code-Install_DC-007ACC?logo=visualstudiocode&logoColor=white)](https://insiders.vscode.dev/redirect/mcp/install?name=dc&config=%7B%22url%22%3A%22https%3A%2F%2Fapi.dynamitecircle.com%2Fmcp%22%7D)
+- **Claude Code:** `claude mcp add --transport http dc https://api.dynamitecircle.com/mcp`
+- **Claude (web / Desktop / mobile):** Settings → Connectors → **Add custom connector** → paste the URL → **Connect**
+- **VS Code:** click the badge above, or `code --add-mcp '{"name":"dc","url":"https://api.dynamitecircle.com/mcp"}'`
+- **Cursor:** click the badge above, or add a custom MCP connector pointing at the URL
+- **ChatGPT / any other MCP app:** add a custom / remote MCP connector pointing at the URL above
+When prompted, **sign in with your DC account** (one-click OAuth) — or send a `dk_` key as `Authorization: Bearer`. The server is always on the current API version, so there's nothing to keep up to date.
+**Hosted MCP vs. this client — which do I want?**
+| You want… | Use |
+|---|---|
+| Zero install, always-current, just MCP tools in a chat app | **Hosted MCP** (URL above) |
+| To call DC from your own Python code (`from dynamitecircle import DC`) | **This client** (library) |
+| A local CLI (`dc profile`) or shell scripting | **This client** (CLI) |
+| A local stdio MCP server (offline-capable, pin a version, dev against localhost) | **This client** (`dc --mcp`) |
+The rest of this README is about **this client**. For the hosted MCP, the URL above is all you need.
 ## Features
 - **Full Member API coverage** — read + write across every public endpoint (run `python3 py/dc.py help` for the live list)
@@ -161,7 +192,18 @@ Run `python3 py/dc.py help` for the full command list.
 ## Setup per AI tool
-### Claude Code
+### Hosted MCP (no install, always current)
+Nothing to clone or update — point your tool at the remote endpoint:
+```bash
+# Claude Code
+claude mcp add --transport http dc https://api.dynamitecircle.com/mcp
+```
+For Claude web / Desktop, Cursor, ChatGPT, and other MCP apps, add a **custom / remote connector** pointing at `https://api.dynamitecircle.com/mcp` (Streamable HTTP). Sign in with your DC account (OAuth) or send a `dk_` key as a Bearer header. Discovery is published at [`/.well-known/mcp.json`](https://api.dynamitecircle.com/.well-known/mcp.json). This is the lowest-effort path; everything below runs the client **locally** instead.
+### Claude Code (local client)
 `.mcp.json` is already shipped with this repo. Open the repo with `claude`:
@@ -178,9 +220,30 @@ pip install -r py/requirements.txt
 Skill discovery (CLI + import) works automatically via `py/SKILL.md`.
+#### Auto-updating local MCP via `uvx` (PyPI)
+Prefer a local stdio server that **pulls the latest published client on every launch** with no clone to maintain? Use [`uv`](https://docs.astral.sh/uv/)'s `uvx` and pass your key in the `env` block:
+```json
+{
+  "mcpServers": {
+    "dc": {
+      "type": "stdio",
+      "command": "uvx",
+      "args": ["--refresh", "--from", "dynamitecircle[mcp]", "dc", "--mcp"],
+      "env": { "DC_API_KEY": "dk_<api-key>" }
+    }
+  }
+}
+```
+Notes: the console script is `dc` (not `dynamitecircle`), so `--from dynamitecircle[mcp]` is required — it both names the package and pulls the optional `mcp` extra. `--refresh` is what makes it auto-update; drop it to let `uvx` cache. The `DC_API_KEY` in `env` wins over any `.env.dc` (an ephemeral `uvx` install has none).
 ### Claude Desktop
-Edit your config file:
+**Easiest — one-click `.mcpb`:** download `dc.mcpb` from the [latest GitHub release](https://github.com/dynamitecircle/dc/releases/latest) and double-click it. Claude Desktop installs the `dc` server and prompts for your `DC_API_KEY` (no JSON editing). Requires Python 3.9+ with the `mcp` package available on your machine (`pip install 'dynamitecircle[mcp]'`). *(Or skip local install entirely and use the [hosted MCP connector](#hosted-mcp-no-install-always-current) above.)*
+**Manual** — edit your config file:
 - macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
 - Linux: `~/.config/Claude/claude_desktop_config.json`
@@ -277,9 +340,19 @@ Non-paginated extras (e.g. `totalUnread` on `inbox`) are passed through under an
 ## Use as a Python library
+Installed from PyPI (`pip install dynamitecircle`) — import by the package name:
+```python
+from dynamitecircle import DC
+dc = DC()
+```
+Working from a clone instead? Point `sys.path` at the `py/` folder and import as `dc`:
 ```python
 import sys
-sys.path.insert(0, "dc")
+sys.path.insert(0, "py")
 from dc import DC
 dc = DC()
@@ -313,8 +386,21 @@ The full live reference for the DC Member API — every endpoint, parameter, and
 The DC Member API ships new endpoints and refinements regularly. This skill is the official client and we update it whenever the API changes. **Plan for updates** — the skill will warn you on stderr the first time a request returns a server `X-API-Version` newer than `DC_API_VERSION`, and major-version bumps may break older clients.
+**The no-maintenance option:** if you use the [hosted MCP](#hosted-mcp-no-install-always-current) (`https://api.dynamitecircle.com/mcp`), there's nothing to update — the server always runs the current API version. The strategies below apply only to the **local client**.
 Pick whichever integration style fits your project. From simplest to most isolated:
+### 0. PyPI — `pip` / `uvx`
+```bash
+pip install dynamitecircle              # CLI + library
+pip install 'dynamitecircle[mcp]'       # + local MCP server (dc --mcp)
+```
+Then `dc setup --api-key dk_<api-key>` and `dc self-test`. Update with `pip install --upgrade dynamitecircle`. For a local MCP that auto-updates on every launch, use the [`uvx` config above](#auto-updating-local-mcp-via-uvx-pypi).
+Best for: the quickest local install, and the auto-updating `uvx` MCP setup.
 ### 1. Plain git clone — quick local use
 ```bash
@@ -453,6 +539,9 @@ dc/
 │
 │   (Future: go/, node/, rs/ folders for sister clients in other languages)
 │
+├── DC/
+│   └── SKILL.md     → ../py/SKILL.md  # human-friendly skill path (symlink)
+│
 ├── docs/                              # ← REAL design docs (canonical)
 │   ├── skill-info.md                # design rules / architecture
 │   └── mcp-info.md                    # MCP setup for every supported client

{dynamitecircle-2.0.1 → dynamitecircle-2.0.4}/docs/mcp-info.md RENAMED Viewed

@@ -1,5 +1,31 @@
 # MCP — Setup for every supported client
+There are **two ways** to get DC's tools into an MCP client:
+1. **Hosted MCP (no install)** — point your client at the remote endpoint `https://api.dynamitecircle.com/mcp`. Nothing to clone, install, or update; always on the current API version. Best for chat apps (Claude web/Desktop, Cursor, ChatGPT).
+2. **Local server (this client)** — run `dc.py --mcp` as a local stdio server with the `--mcp` flag. Best when you want offline capability, a pinned version, or to develop against a local API. Most of this doc covers this path.
+## Hosted MCP (no install)
+`https://api.dynamitecircle.com/mcp` (Streamable HTTP). Authenticate with one-click OAuth (reuses your DC login) or a `dk_` API key as a Bearer header.
+| Client | How |
+|---|---|
+| **Claude Code** | `claude mcp add --transport http dc https://api.dynamitecircle.com/mcp` |
+| **Claude web / Desktop / mobile** | Settings → Connectors → **Add custom connector** → paste URL → **Connect** → sign in with DC (paid plan; Team/Enterprise admins may need to approve first) |
+| **VS Code** | `code --add-mcp '{"name":"dc","url":"https://api.dynamitecircle.com/mcp"}'` (or the badge below) |
+| **Cursor / ChatGPT / other MCP apps** | Add a custom / remote MCP connector pointing at the URL above |
+One-click badges (hosted MCP):
+[![Add to Cursor](https://cursor.com/deeplink/mcp-install-dark.svg)](cursor://anysphere.cursor-deeplink/mcp/install?name=dc&config=eyJ1cmwiOiJodHRwczovL2FwaS5keW5hbWl0ZWNpcmNsZS5jb20vbWNwIn0=)
+&nbsp;
+[![Install in VS Code](https://img.shields.io/badge/VS_Code-Install_DC-007ACC?logo=visualstudiocode&logoColor=white)](https://insiders.vscode.dev/redirect/mcp/install?name=dc&config=%7B%22url%22%3A%22https%3A%2F%2Fapi.dynamitecircle.com%2Fmcp%22%7D)
+Discovery metadata is published at [`/.well-known/mcp.json`](https://api.dynamitecircle.com/.well-known/mcp.json); the OAuth protected-resource descriptor is at `/.well-known/oauth-protected-resource`. The server is also listed in the [Official MCP Registry](https://registry.modelcontextprotocol.io) as `io.github.dynamitecircle/dc` (which feeds VS Code, Smithery, PulseMCP, and others).
+## Local server (this client)
 The `dc` skill exposes its commands as [Model Context Protocol](https://modelcontextprotocol.io) tools when run with the `--mcp` flag. The protocol is standard, so the same server works in every MCP-compatible client.
 ## One-time install
@@ -22,6 +48,31 @@ python3 py/dc.py --mcp
 The server speaks JSON-RPC over stdio. It's not meant to be used directly — wire it into a client below.
+## Auto-updating local server via `uvx` (PyPI)
+To run a local stdio server **without a clone** that pulls the latest published client on every launch, use [`uv`](https://docs.astral.sh/uv/)'s `uvx`:
+```json
+{
+  "mcpServers": {
+    "dc": {
+      "type": "stdio",
+      "command": "uvx",
+      "args": ["--refresh", "--from", "dynamitecircle[mcp]", "dc", "--mcp"],
+      "env": { "DC_API_KEY": "dk_<api-key>" }
+    }
+  }
+}
+```
+Three things make this work:
+- **`--from dynamitecircle[mcp]`** — the package is `dynamitecircle` but the console script is `dc`, so you can't write `uvx dynamitecircle`; `--from` names the package (and the `[mcp]` extra installs the MCP dependency) while `dc` is the command to run.
+- **`--refresh`** — forces `uvx` to check PyPI for a newer version instead of reusing its cache. Drop it if you'd rather pin to whatever's cached.
+- **`env.DC_API_KEY`** — an ephemeral `uvx` install has no `py/.env.dc`, so pass the key via the env block. It takes precedence over any `.env.dc` anyway (existing env vars always win).
+Same idea with `pipx`: `pipx run --spec 'dynamitecircle[mcp]' dc --mcp` (add `--no-cache` to force the latest).
 ## Quick check (no client needed)
 ```bash
@@ -129,6 +180,24 @@ mcp_servers:
 `GEMINI.md` (symlinked to `CLAUDE.md`) provides additional context for the assistant.
+### VS Code
+VS Code has native MCP support and reads `.vscode/mcp.json` (per-workspace) or your user `mcp.json`. For the **hosted** server, the one-click badge above or `code --add-mcp '{"name":"dc","url":"https://api.dynamitecircle.com/mcp"}'` is all you need. For a **local** stdio server, add to `.vscode/mcp.json`:
+```json
+{
+  "servers": {
+    "dc": {
+      "type": "stdio",
+      "command": "python3",
+      "args": ["/absolute/path/to/dc/py/dc.py", "--mcp"]
+    }
+  }
+}
+```
+VS Code shows inline **Start / Stop / Restart** actions above each server in that file. It also renders the [MCP registry](https://code.visualstudio.com/docs/agent-customization/mcp-servers) in the Extensions view, so once `dc` is listed there it's installable from inside the editor.
 ### Cursor
 1. Open Settings → MCP
@@ -138,6 +207,8 @@ mcp_servers:
    - Command: `python3`
    - Args: `/absolute/path/to/dc/py/dc.py --mcp`
+Or use the one-click **Add to Cursor** badge above (hosted MCP).
 ### Cline / Continue / Windsurf / Zed
 Same shape. Each tool has its own settings UI but the data is the same:

{dynamitecircle-2.0.1 → dynamitecircle-2.0.4}/docs/skill-info.md RENAMED Viewed

@@ -77,11 +77,25 @@ Fields:
 | Field | Required | Meaning |
 |---|---|---|
-| `name` | MUST | Public command name. Kebab-case. Used in CLI and as MCP tool name |
-| `help` | SHOULD | One-line description for `help` listing |
+| `name` | MUST | Public command name. **Kebab-case** — used verbatim as the CLI command (`dc trip-create`) and converted to snake_case for the MCP tool name (`mcp__dc__trip_create`) |
+| `help` | SHOULD | One-line description for `help` listing (also the MCP tool description) |
 | `parser` | optional | Static method on `_DCCore` that converts raw CLI args → `(args, kwargs)`. Required if the command takes flags |
-There are **no** `mode="read"|"write"` labels, no permission gates, no extension hooks. The decorator is intentionally minimal.
+The decorator carries **no** `mode="read"|"write"` label, no permission gates, no extension hooks — it stays minimal. Read/write classification for MCP annotations is derived centrally instead (see Tool naming below).
+## Tool naming and casing
+Three surfaces, three casings — each idiomatic for its context. Don't "fix" them to match each other:
+| Surface | Form | Example |
+|---|---|---|
+| CLI command | kebab-case | `dc trip-create`, `--start-date` |
+| Python method / library | snake_case | `dc.trip_create()` |
+| MCP tool + fields | snake_case | `mcp__dc__trip_create`, field `event_id` |
+The MCP layer converts kebab → snake at the boundary (`run_mcp` for tool names, `_build_input_schema` for fields); the CLI parser still gets kebab. Tool names are **noun-first** (`trip_create`, not `create_trip`). This is deliberate: alphabetical tool lists then cluster all operations on the same resource together (`trip_create` / `trip_delete` / `trip_update` / `trips` sit adjacent), which is more discoverable for an agent scanning `tools/list` than a verb-first scheme that scatters them by action. A verb-first reorder was considered and **rejected** for that reason (and to avoid breaking renames across all three surfaces).
+**Annotations:** each MCP tool advertises `readOnlyHint` / `destructiveHint` / `openWorldHint`. The read/write split lives in `_WRITE_COMMANDS` (in `dc.py`), derived from each command's actual HTTP verb and guarded by `tests/test_annotations.py`, which re-derives the set from source and fails on drift — so a new write command can't silently ship mislabelled as read-only.
 ## Argument parsing
@@ -221,8 +235,8 @@ How it works:
 1. `_MCP_AVAILABLE` flag set at module load via `try: from mcp.server import Server`
 2. `--mcp` in `sys.argv` → entry point calls `DC().run_mcp()`
-3. `run_mcp()` builds an `mcp.server.Server`, registers `list_tools` (one per `@skill_command`) and `call_tool` (translates MCP args → CLI raw_args, then invokes via `self._invoke`)
-4. Result serialized as `TextContent` (JSON for dicts/lists, str for scalars)
+3. `run_mcp()` builds an `mcp.server.Server`, registers `list_tools` (one per `@skill_command`, each with a snake_case name + `readOnlyHint`/`destructiveHint`/`openWorldHint` annotations) and `call_tool` (translates MCP args → CLI raw_args, then invokes via `self._invoke`)
+4. Result serialized as `TextContent` (JSON for dicts/lists, str for scalars); object results are *also* returned as `structuredContent` for structure-aware clients
 If `mcp` is not installed → clean install hint, no traceback. CLI and Python-import users have **zero** dependencies.
@@ -231,9 +245,10 @@ If `mcp` is not installed → clean install hint, no traceback. CLI and Python-i
 1. Add the business logic to `_DCCore` (returns plain Python data)
 2. If it takes flags → add a `_parse_*` static method to `_DCCore`
 3. Add a `@skill_command`-decorated wrapper to `DC` that delegates to `_core`
-4. Update `SKILL.md` with usage examples
-5. Verify the new endpoint is documented in the live API reference: https://www.dynamitecircle.com/developers/
-6. Verify:
+4. **If it mutates (POST/PATCH/PUT/DELETE), add its name to `_WRITE_COMMANDS`** in `dc.py` — `tests/test_annotations.py` re-derives writes from the HTTP verbs and fails if you forget, so a write can't ship mislabelled as a safe read
+5. Update `SKILL.md` with usage examples
+6. Verify the new endpoint is documented in the live API reference: https://www.dynamitecircle.com/developers/
+7. Verify:
    ```bash
    python3 py/dc.py help                    # appears in list
    python3 py/dc.py <new-command> --json    # round-trips
@@ -246,7 +261,7 @@ These were design choices. **Do not add them without a discussion**:
 - ❌ Telemetry / usage tracking
 - ❌ Error logging to disk (`.errors.jsonl`)
-- ❌ Read/write mode labels on commands
+- ❌ Read/write **gates** (permission enforcement) — note: read/write *annotations* for MCP clients DO exist (advisory hints derived from HTTP verbs), but nothing in the client blocks a write
 - ❌ Extension hooks (`@skill_extension`)
 - ❌ Owner-binding / runtime metadata
 - ❌ Lint rules / sweep / pulse

{dynamitecircle-2.0.1 → dynamitecircle-2.0.4}/manifest.json RENAMED Viewed

@@ -2,7 +2,7 @@
   "manifest_version": "0.3",
   "name": "dc",
   "display_name": "DC Member API",
-  "version": "2.0.1",
+  "version": "2.0.4",
   "description": "Read and act on your own Dynamite Circle membership data via the public Member API.",
   "long_description": "Single-file Python **client** for the public DC Member API at https://api.dynamitecircle.com. Same file runs three ways: CLI, Python library, and MCP server (--mcp flag, opt-in mcp dependency). Compatible with Claude Code, Claude Desktop, Codex CLI, Gemini CLI, Cursor, GitHub Copilot, and any other Agent Skills / MCP-compatible tool. Covers profile (read+update), trips (CRUD), events + RSVP, virtual events + RSVP, tickets, invites, inbox, rooms, chapters, places lookup, locator digest, announcements, and event-day agendas (DCBKK / DCMEX / DC BLACK schedules, meetups, sponsors, session bookmarks).",
   "author": {

dynamitecircle 2.0.1__tar.gz → 2.0.4__tar.gz

dynamitecircle 2.0.1tar.gz → 2.0.4tar.gz