Skip to content

urveshdev/tradelane-intel

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Tradelane Intel — ocean container-shipping analyst

An ocean container-shipping analyst you can hand any question about a tradelane, port, service, or carrier network — and get a grounded, cited answer built from live public data (alliance network PDFs, carrier newsrooms, freight-rate indices, port-authority stats, reputable trade press).

Scope: container liner shipping only — the ocean carriers, alliances, services, ports, and freight rates that move the world's boxes. Not air freight, trucking/rail, or bulk/tanker shipping.

⚠️ Disclaimer — verify before you act. This is an analytical aid, not an authoritative schedule or booking source. Answers are reconstructed from public data that can be dated, incomplete, or superseded, and AI can misread a rotation. Always confirm exact service details, port rotations, transit times, cut-offs, and rates directly with the carrier (or your forwarder) before making any commercial, routing, or booking decision, and do your own research. Freight-rate figures are third-party market commentary, not quotes. No warranty; use at your own risk.

It ships no dataset. Every answer is rebuilt at question time from public sources, so it's always current and independently checkable. What it gives you is the method: where the answer lives, how to fetch it (and recover when a carrier site blocks you), the analytical frame that turns a schedule into an insight, and the discipline to cite everything and never fabricate.


How to use it

▸ Claude — install the plugin (Claude app UI or Claude Code)

This repo is a plugin marketplace, so it installs as a plugin in both the Claude app and Claude Code. The marketplace is the public GitHub repo: urveshdev/tradelane-intel. (Steps below are the current official flows — see the cited Anthropic help articles.)

Claude app — no terminal (Pro, Max, Team, or Enterprise plan):

  1. Open Customize → Plugins tab.
  2. In Personal plugins, click "+"Add marketplaceAdd from a repository.
  3. Point it at the GitHub repo urveshdev/tradelane-intel (or the repo URL) to sync the marketplace.
  4. Open the synced marketplace and install the tradelane-intel plugin.

Steps per Anthropic's Use plugins in Claude.

Claude Code — terminal:

/plugin marketplace add urveshdev/tradelane-intel
/plugin install tradelane-intel@tradelane-intel
/reload-plugins

/reload-plugins activates it without a restart. Update later with /plugin marketplace update tradelane-intel. Steps per Anthropic's Discover and install plugins.

Either way, ask a tradelane / port / carrier question and it kicks in.

▸ ChatGPT, Gemini, Copilot, or a quick Claude chat — copy one file (no install)

Fastest, works anywhere with browsing:

  1. Open the prompt → view it / raw text.
  2. Select all → copy → paste as your first message in a new chat, turn on web browsing, ask.

Reuse it without re-pasting by dropping that text into a Project's custom instructions (Claude.ai or ChatGPT). For a shareable ChatGPT Custom GPT, follow dist/custom-gpt-setup.md.

Requirements (any path): the assistant's web browsing/search turned on. No API keys, no dataset, no build step. (Python is only needed for the optional HHI helper script.)


What's in the package

tradelane-intel/
  README.md                                    this file
  LICENSE                                      MIT
  .claude-plugin/marketplace.json              makes this repo a Claude Code plugin marketplace
  plugins/tradelane-intel/
    .claude-plugin/plugin.json                 the plugin manifest
    skills/tradelane-intel/                    the skill Claude loads
      SKILL.md                                   entry point — workflow + non-negotiables
      references/                                the knowledge (also uploadable to a ChatGPT Custom GPT)
        insight-playbook.md                        question patterns → frame → good-answer shape
        method.md                                  the 6 analytical frames + the editorial standard
        carrier-sources.md                         where carriers/alliances/ports publish — with URLs
        rates-and-demand.md                        rate indices + demand sources — with URLs
        fetching.md                                how to fetch; bot-wall & map-PDF recovery
      scripts/concentration.py                     HHI + top-3 from counts you gather (no data inside)
  dist/                                        for ChatGPT / Gemini / any assistant with browsing
    tradelane-intel.txt                          ← the single copy-paste prompt (self-contained)
    portable-prompt.md                           same prompt in markdown, with usage notes
    custom-gpt-setup.md                          step-by-step to build a ChatGPT Custom GPT

The two paths use the same content — the Claude skill reads the references/ files; the .txt inlines those same references into one self-contained, paste-able prompt.


What it can answer (examples, not limits)

  • "Who runs Asia–West Africa, and what's the rotation?"
  • "Is [port] a transshipment hub or a gateway — and how would I know?"
  • "Why are rates on my lane moving — and is it demand or capacity discipline?"
  • "How concentrated is the transpacific vs Asia–Europe, and who has pricing power?"
  • "What ports does [carrier]'s [service] call, and how long is the transit?"

The principles it enforces

  • Cite every fact; separate carrier-published from third-party market commentary.
  • Never fabricate a rotation or a rate — say "not found" instead.
  • "New" needs this-quarter traction, not a launch date; otherwise it's "confirming a trend".
  • Test hunches, don't just confirm them — refuting a wrong hunch with data is the best answer.
  • Match depth to the question; one proven point beats five shallow ones.

No scraped data is distributed. All sources are public and cited per answer.

About

Ocean container-shipping tradelane analyst — answer any tradelane/port/service/carrier question from live public data with citations. Portable to Claude, ChatGPT, Gemini. No dataset bundled.

Topics

Resources

License

Stars

1 star

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages