Add trade-intelligence skill (draft until MCP tools ship)

.claude-plugin/marketplace.json

@@ -4,15 +4,15 @@
     "name": "Terminal49"
   },
   "metadata": {
-    "description": "Terminal49 container-tracking tools and agent skills",
-    "version": "0.1.0"
+    "description": "Terminal49 container-tracking and trade-intelligence tools and agent skills",
+    "version": "0.2.0"
   },
   "plugins": [
     {
       "name": "terminal49",
       "source": "./plugins/terminal49",
-      "description": "Use Terminal49 MCP tools to track and investigate container shipments",
-      "version": "0.1.0"
+      "description": "Use Terminal49 MCP tools to track and investigate container shipments, and Terminal49 trade intelligence to research US ocean imports",
+      "version": "0.2.0"
     }
   ]
 }

.cursor-plugin/marketplace.json

@@ -4,15 +4,15 @@
     "name": "Terminal49"
   },
   "metadata": {
-    "description": "Terminal49 container-tracking tools and agent skills",
-    "version": "0.1.0"
+    "description": "Terminal49 container-tracking and trade-intelligence tools and agent skills",
+    "version": "0.2.0"
   },
   "plugins": [
     {
       "name": "terminal49",
       "source": "./plugins/terminal49",
-      "description": "Use Terminal49 MCP tools to track and investigate container shipments",
-      "version": "0.1.0"
+      "description": "Use Terminal49 MCP tools to track and investigate container shipments, and Terminal49 trade intelligence to research US ocean imports",
+      "version": "0.2.0"
     }
   ]
 }

AGENTS.md

@@ -12,11 +12,19 @@ This is a public, cross-platform plugin marketplace for Terminal49.
   resource identifier.
 - Do not ask users to paste credentials into chat. Let the client initiate its
   OAuth connection flow.
+- The `trade-intelligence` skill reads a Terminal49 API key that the user saves
+  in their own terminal (`T49_API_KEY` or `~/.t49_api_key`). Never put a real
+  key in examples, logs, or chat.
 
 ## Shared source of truth
 
-- `plugins/terminal49/skills/container-tracking/SKILL.md` is the shared behavioral
-  guidance for all clients.
+- The skills under `plugins/terminal49/skills/` are the shared behavioral
+  guidance for all clients:
+  - `container-tracking/SKILL.md` covers container and shipment questions with
+    the Terminal49 MCP tools.
+  - `trade-intelligence/SKILL.md` covers US import market research through the
+    Terminal49 REST API, using its bundled `scripts/ti.py` client and
+    `references/api.md`.
 - `plugins/terminal49/.mcp.json` and `plugins/terminal49/mcp.json` are the MCP
   adapters for Claude/Codex and Cursor respectively. Keep their server entries
   identical.

CONTRIBUTING.md

@@ -4,7 +4,7 @@ Thanks for improving the Terminal49 agent plugins.
 
 ## Principles
 
-1. Keep the shared skill client-neutral.
+1. Keep the shared skills client-neutral.
 2. Prefer tool-selection guidance over duplicating the Terminal49 API docs.
 3. Treat `track_container` as a write: search first and create a tracking
    request only when the user intends to start tracking.
@@ -20,4 +20,4 @@ Thanks for improving the Terminal49 agent plugins.
 4. Test installation in any client affected by the change.
 
 Platform-specific behavior belongs in its adapter. Guidance that applies to all
-clients belongs in `plugins/terminal49/skills/container-tracking/SKILL.md`.
+clients belongs in the relevant skill under `plugins/terminal49/skills/`.

README.md

@@ -1,12 +1,14 @@
 # Terminal49 agent plugins
 
-Install Terminal49 container-tracking tools and workflow guidance in Grok Bot,
-Cursor, Claude Code, Codex, and GitHub Copilot CLI.
+Install Terminal49 container-tracking tools, trade-intelligence research, and
+workflow guidance in Grok Bot, Cursor, Claude Code, Codex, and GitHub Copilot
+CLI.
 
 This marketplace currently publishes one plugin:
 
 - **Terminal49** — connects to the hosted Terminal49 MCP server and teaches
-  agents how to search, track, and investigate container shipments.
+  agents how to search, track, and investigate container shipments, and how to
+  research US ocean imports with Terminal49 trade intelligence.
 
 ## What the plugin provides
 
@@ -16,9 +18,14 @@ This marketplace currently publishes one plugin:
   demurrage-risk questions
 - Guardrails around tracking-request creation, credentials, dates, and missing
   data
+- A `trade-intelligence` skill for competitor research, sourcing, and market
+  trends from US ocean import bill-of-lading records since 2022 (for accounts
+  with trade intelligence enabled)
 
-The plugin does not contain API keys or customer data. Authentication happens
-through the MCP client's OAuth flow.
+The plugin does not contain API keys or customer data. Container tracking
+authenticates through the MCP client's OAuth flow. Trade intelligence currently
+uses a Terminal49 API key that you save in your own terminal; see the
+[plugin README](plugins/terminal49/README.md#trade-intelligence-api-key).
 
 ## Install
 
@@ -77,9 +84,12 @@ After connecting your Terminal49 account, ask your agent:
 - “Is this container ready for pickup, and are there any holds?”
 - “Explain what changed in this container's journey.”
 - “Show containers updated since yesterday and group them by status.”
+- “Who are the largest US importers of office chairs, and where do they source
+  them?”
 
 The example identifier is illustrative. Results depend on the shipments visible
-to the authenticated Terminal49 account.
+to the authenticated Terminal49 account. Trade-intelligence questions need an
+account with trade intelligence enabled.
 
 ## Repository layout
 
@@ -96,10 +106,15 @@ to the authenticated Terminal49 account.
     ├── mcp.json                      # Cursor MCP adapter
     ├── README.md
     ├── assets/logo.svg               # Referenced by the Cursor manifest
-    └── skills/container-tracking/SKILL.md
+    └── skills/
+        ├── container-tracking/SKILL.md
+        └── trade-intelligence/
+            ├── SKILL.md
+            ├── references/api.md     # Endpoint parameters and fields
+            └── scripts/ti.py         # Python 3 API client, no dependencies
 ```
 
-The skill is shared. Each platform-specific manifest and MCP file is a thin
+The skills are shared. Each platform-specific manifest and MCP file is a thin
 adapter around that shared content, with validation keeping the endpoint in
 sync.
 

plugins/terminal49/.claude-plugin/plugin.json

@@ -2,8 +2,8 @@
   "$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
   "name": "terminal49",
   "displayName": "Terminal49",
-  "version": "0.1.0",
-  "description": "Track and investigate container shipments with Terminal49 MCP tools and workflow guidance.",
+  "version": "0.2.0",
+  "description": "Track and investigate container shipments with Terminal49 MCP tools, and research US ocean import markets with Terminal49 trade intelligence.",
   "author": {
     "name": "Terminal49"
   },
@@ -15,7 +15,8 @@
     "logistics",
     "mcp",
     "ocean-freight",
-    "supply-chain"
+    "supply-chain",
+    "trade-intelligence"
   ],
   "skills": "./skills/",
   "mcpServers": "./.mcp.json"

plugins/terminal49/.codex-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "terminal49",
-  "version": "0.1.0",
-  "description": "Track and investigate container shipments with Terminal49 MCP tools and workflow guidance.",
+  "version": "0.2.0",
+  "description": "Track and investigate container shipments with Terminal49 MCP tools, and research US ocean import markets with Terminal49 trade intelligence.",
   "author": {
     "name": "Terminal49"
   },
@@ -13,14 +13,15 @@
     "logistics",
     "mcp",
     "ocean-freight",
-    "supply-chain"
+    "supply-chain",
+    "trade-intelligence"
   ],
   "skills": "./skills/",
   "mcpServers": "./.mcp.json",
   "interface": {
     "displayName": "Terminal49",
-    "shortDescription": "Track and investigate container shipments",
-    "longDescription": "Connect to Terminal49 and use guided workflows for container status, pickup readiness, holds, ETAs, routes, delays, and demurrage risk.",
+    "shortDescription": "Track container shipments and research US imports",
+    "longDescription": "Connect to Terminal49 and use guided workflows for container status, pickup readiness, holds, ETAs, routes, delays, and demurrage risk, plus trade-intelligence research on US ocean imports: competitors, sourcing, and market trends.",
     "developerName": "Terminal49",
     "category": "Productivity",
     "capabilities": [
@@ -29,7 +30,8 @@
     ],
     "websiteURL": "https://terminal49.com",
     "defaultPrompt": [
-      "Track a container or investigate the latest status of my shipments"
+      "Track a container or investigate the latest status of my shipments",
+      "Show what a competitor imports, from where, and how its volume is trending"
     ]
   }
 }

plugins/terminal49/.cursor-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "terminal49",
   "displayName": "Terminal49",
-  "version": "0.1.0",
-  "description": "Track and investigate container shipments with Terminal49 MCP tools and workflow guidance.",
+  "version": "0.2.0",
+  "description": "Track and investigate container shipments with Terminal49 MCP tools, and research US ocean import markets with Terminal49 trade intelligence.",
   "author": {
     "name": "Terminal49"
   },
@@ -15,7 +15,8 @@
     "logistics",
     "mcp",
     "ocean-freight",
-    "supply-chain"
+    "supply-chain",
+    "trade-intelligence"
   ],
   "category": "Productivity",
   "skills": "./skills/",

plugins/terminal49/README.md

@@ -1,7 +1,8 @@
 # Terminal49 plugin
 
-This plugin connects an agent to Terminal49's hosted MCP server and supplies a
-shared skill for reliable container-tracking workflows. It is published from
+This plugin connects an agent to Terminal49's hosted MCP server and supplies
+shared skills for reliable container-tracking workflows and trade-intelligence
+research. It is published from
 the [Terminal49/agent-plugins](https://github.com/Terminal49/agent-plugins)
 marketplace — see that README for installation instructions.
 
@@ -11,7 +12,22 @@ The MCP client should open an OAuth authorization flow when Terminal49 tools are
 first used. Authenticate in the browser and return to the client. Do not paste
 API keys, access tokens, or one-time codes into chat.
 
-## Included skill
+### Trade intelligence API key
+
+The `trade-intelligence` skill calls the Terminal49 REST API with your
+Terminal49 API key and needs Python 3. Create a key at
+https://app.terminal49.com/developers/api-keys, then save it in your own
+terminal (bash or zsh). Never paste it into chat:
+
+```sh
+printf 'Terminal49 API key: '; read -rs K; echo; (umask 077; touch ~/.t49_api_key; chmod 600 ~/.t49_api_key; printf %s "$K" > ~/.t49_api_key); unset K
+```
+
+Alternatively, set `T49_API_KEY` in the environment your agent runs in.
+Trade intelligence is enabled per account. If requests return HTTP 403, contact
+[email protected].
+
+## Included skills
 
 `container-tracking` is written for logistics operators and covers:
 
@@ -27,11 +43,22 @@ API keys, access tokens, or one-time codes into chat.
 - querying fleets with filters and pagination
 - presenting dates, holds, missing information, and paid-feature boundaries
 
+`trade-intelligence` answers market questions from US ocean import
+bill-of-lading records since 2022, refreshed daily:
+
+- competitor research: what a named company imports, from which countries,
+  through which US ports and carriers, and how its volume is trending
+- sourcing: who imports a product and where it comes from
+- market trends by port, coast, carrier, origin, or commodity
+- caveats that change the answer: forwarders and brokers on the records, one
+  company under several names, importers who withhold their names, partial
+  months, and estimated values
+
 The Terminal49 API and MCP implementation live in
 [Terminal49/API](https://github.com/Terminal49/API).
 
 ## Plugin evals
 
-Behavioral eval cases use checked-in MCP mocks and never require the live
-Terminal49 service. See [`evals/README.md`](evals/README.md) for validation,
+Behavioral eval cases cover `container-tracking`, use checked-in MCP mocks,
+and never require the live Terminal49 service. See [`evals/README.md`](evals/README.md) for validation,
 mocked eval, baseline, and marketplace-submission guidance.