Ask Codex for the service's own OAuth scopes

.claude-plugin/marketplace.json

@@ -9,7 +9,7 @@
       "name": "nutrient",
       "source": "./plugins/nutrient",
       "description": "Parse, extract from and classify documents with Nutrient DWS: OCR for scans, Markdown with layout, structured fields with citations, and labels, as stored runs in your Nutrient account.",
-      "version": "1.0.0",
+      "version": "1.0.1",
       "author": { "name": "Nutrient", "url": "https://www.nutrient.io" },
       "homepage": "https://www.nutrient.io/api/",
       "license": "MIT"

AGENTS.md

@@ -22,6 +22,10 @@ OpenAI's directory takes a ZIP of `plugins/nutrient`, uploaded for each release.
 
 Never change `name` in `plugin.json`: it is the plugin's permanent identity.
 
+Keep `scopes` in `.mcp.json`. Without it, Codex asks for every scope the authorization server lists instead of the
+service's own, and the service can't exchange that grant, so every tool call fails. Claude Code asks for the same pair,
+taken from the service's sign-in challenge, and ChatGPT for the scopes the service advertises. `ci/check.py` checks it.
+
 ## Refreshing the tool snapshot
 
 `ci/tools.json` and `ci/tools-chatgpt.json` are the catalogs Hosted MCP (`hosted-mcp/` in the private Nutrient monorepo)

ci/check.py

@@ -133,6 +133,9 @@ def collect(node, names):
 servers = load(PLUGIN / ".mcp.json").get("mcpServers", {})
 check(list(servers) == ["nutrient"] and servers["nutrient"].get("type") == "http" and servers["nutrient"].get("url") == SERVICE,
       f".mcp.json declares one server, nutrient, at {SERVICE} over HTTP (OpenAI connects one server per plugin)")
+# Without scopes, Codex asks for every scope the authorization server lists, and the service can't use that grant.
+check(servers.get("nutrient", {}).get("scopes") == ["mcp:invoke", "offline_access"],
+      ".mcp.json requests the service's own scopes, mcp:invoke and offline_access")
 
 skills = sorted((PLUGIN / "skills").glob("*/SKILL.md"))
 check(len(skills) > 0, "the plugin has at least one skill")

plugins/nutrient/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "nutrient",
   "displayName": "Nutrient",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Parse, extract from and classify documents with Nutrient DWS: OCR for scans, Markdown with layout, structured fields with citations, and labels, as stored runs in your Nutrient account.",
   "author": { "name": "Nutrient", "url": "https://www.nutrient.io" },
   "homepage": "https://www.nutrient.io/api/",

plugins/nutrient/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "nutrient",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Parse, extract from and classify documents with Nutrient DWS: OCR for scans, Markdown with layout, structured fields with citations, and labels, as stored runs in your Nutrient account.",
   "author": {
     "name": "Nutrient",

plugins/nutrient/.mcp.json

@@ -2,7 +2,11 @@
   "mcpServers": {
     "nutrient": {
       "type": "http",
-      "url": "https://mcp.nutrient.io/mcp"
+      "url": "https://mcp.nutrient.io/mcp",
+      "scopes": [
+        "mcp:invoke",
+        "offline_access"
+      ]
     }
   }
 }

plugins/nutrient/README.md

@@ -7,9 +7,12 @@ file is attached to the chat.
 
 ## Use it
 
-1. Add the plugin, then connect the Nutrient connector and sign in with your Nutrient account. In Claude Code, install it
-   with `/plugin install nutrient --marketplace PSPDFKit-labs/nutrient-mcp-plugins`; in Codex, with
-   `codex plugin marketplace add PSPDFKit-labs/nutrient-mcp-plugins` and `codex plugin add nutrient@nutrient`.
+1. Add the plugin, then connect the Nutrient connector and sign in with your Nutrient account. The tools appear once
+   you're signed in.
+   - In Claude Code, install it with `/plugin install nutrient --marketplace PSPDFKit-labs/nutrient-mcp-plugins`, then
+     run `/mcp`, choose `plugin:nutrient:nutrient` and sign in.
+   - In Codex, install it with `codex plugin marketplace add PSPDFKit-labs/nutrient-mcp-plugins` and
+     `codex plugin add nutrient@nutrient`, sign in with `codex mcp login nutrient`, then start a new session.
 2. Attach a PDF, Office document or image, point to a file on disk or an HTTPS link, and ask to parse it, convert it to
    Markdown, extract fields or tables, or classify it.
 3. The assistant sends the file to your Nutrient account and returns the result: Markdown with layout, structured fields