Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/OrcaQubits/agentic-commerce-skills-pluginsnpx agentmods add skills/orcaqubits/agentic-commerce-skills-plugins/nlweb-chatgpt-appsdkWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/orcaqubits/agentic-commerce-skills-plugins/nlweb-chatgpt-appsdk)<a href="https://agentmods.dev/skills/orcaqubits/agentic-commerce-skills-plugins/nlweb-chatgpt-appsdk"><img src="https://agentmods.dev/badge/skills/orcaqubits/agentic-commerce-skills-plugins/nlweb-chatgpt-appsdk/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/orcaqubits/agentic-commerce-skills-plugins/nlweb-chatgpt-appsdk"><img src="https://agentmods.dev/badge/skills/orcaqubits/agentic-commerce-skills-plugins/nlweb-chatgpt-appsdk.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00103 | $0.01828 |
| Opus 5 | $0.00051 | $0.00914 |
| Sonnet 5 | $0.00021 | $0.00366 |
| Haiku 4.5 | $0.00010 | $0.00183 |
Grade A, and why
nlweb-chatgpt-appsdk scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.
NLWeb + ChatGPT Apps SDK
Before writing code
Fetch live docs:
- Fetch https://github.com/nlweb-ai/NLWeb/blob/main/docs/nlweb-chatgpt-integration.md for the canonical wiring steps.
- Fetch https://github.com/nlweb-ai/NLWeb/blob/main/docs/nlweb-appsdk-adapter.md for the port-8100 adapter.
- Inspect
openai-apps-sdk-integration/nlweb_server_node/in the live repo — this is a separate Node.js/TypeScript MCP server. - Web-search OpenAI's current Apps SDK documentation — link from https://platform.openai.com.
- Verify the React widget contract — what
nlweb-list.htmlexpects from the tool result envelope.
Conceptual Architecture
Why a Separate Adapter?
NLWeb's native /mcp returns generic MCP tool results — a list of Schema.org objects. ChatGPT's Apps SDK expects a richer envelope that pairs tool output with a UI resource (widget). The integration ships two extra pieces on top of the base NLWeb server:
-
openai-apps-sdk-integration/nlweb_server_node/— a Node.js/TypeScript MCP server that registers annlweb-listtool and serves a React widget atui://widget/nlweb-list.html. ChatGPT discovers this and renders the widget when the tool returns results. -
AppSDK Adapter (port 8100,
appsdk_adapter_server.py) — an aiohttp server that takes legacy NLWeb message-list responses and re-shapes them into the OpenAI Apps SDK envelope on the fly. Useful if you already have NLWeb running and want to plug ChatGPT in without rewriting clients.
Native MCP clients (Claude, Gemini) ignore both — they hit the canonical /mcp on port 8000.
The Two Integration Paths
| Path | Use When |
|---|---|
| Node.js MCP server with widget | You want the full ChatGPT app experience with the React widget |
| Port-8100 AppSDK adapter | You only want translation; you don't need a widget |
For a new build, use path 1. For retrofitting an existing NLWeb deployment, path 2 is lower-risk.
The nlweb-list Tool
The Node.js server registers a single MCP tool, nlweb-list, with a schema like:
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 162 lines · 103 tokens per session scan A f14595450475
nlweb-chatgpt-appsdk is a skill published in the GitHub repository OrcaQubits/agentic-commerce-skills-plugins (39 stars, last pushed 2d ago), licensed MIT. It adds 103 tokens to every session and 1,828 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-15.
Other skills, from other repositories
archestra-dev-frontend
Use when modifying Archestra frontend Next.js/React code, UI components, forms, TanStack Query hooks, generated API client usage, frontend copy, or documentation links.
dflow-phantom-connect
Build Solana wallet-connected apps with Phantom Connect SDKs and DFlow spot trading. Use when user asks to connect a Phantom wallet, integrate Phantom in React, React Native, or vanilla JS, sign messages or transactions, build token-gated pages, mint NFTs, accept crypto payments, or swap/stream tokens with DFlow.…
near-dapp
Build NEAR Protocol dApps. Use for: (1) creating new NEAR dApps with create-near-app (Vite+React, Next.js), (2) adding NEAR wallet connection to existing apps with @hot-labs/near-connect and near-connect-hooks, (3) building frontend UI for NEAR smart contracts, (4) integrating wallet sign-in/sign-out, contract calls…
codegen-react
Scaffold and iterate on Vite + React applications.
design-taste-frontend
Use for visual design direction on greenfield, user-facing surfaces — marketing and landing pages, generated apps, portfolio-style or standalone pages. NOT for routine Archestra platform UI work (dashboards, data tables, settings, forms, existing components) — use archestra-dev-frontend for that. Upstream intent…
algolia-search-v2
Algolia Search Integration workflow skill. Use this skill when the user needs Expert patterns for Algolia search implementation, indexing and the operator should preserve the upstream workflow, copied support files, and provenance before merging or handing off.