NexScope Amazon Intelligence

简体中文

An OpenAI Codex plugin containing 40 Amazon ecommerce skills for product, market, keyword, competitor, review, sales, policy, category, and advertising workflows.

Contents

• .codex-plugin/plugin.json — plugin manifest and public listing metadata
• skills/ — 40 independently loadable Amazon skills
• assets/ — plugin icon and light/dark logos
• SUBMISSION.md — public listing copy, test cases, and publisher-side release requirements

The skills are synchronized from nexscope-ai/nexscope-ecommerce-skills at commit e380b29273c2d1cd77e3335c2979c6310f484db4. The two upstream skills absent from the earlier local snapshot, amazon-ads-manager and amazon-category-lookup, are included.

Authentication

Paid NexScope subscribers automatically receive plugin access. The browser device flow verifies the subscription, binds the installation to the subscriber's existing API key, and stores only a revocable installation token in macOS Keychain or Windows Credential Manager:

python3 runtime/installer.py connect
python3 runtime/installer.py update

All skill scripts load the connected device credential first. Developers can explicitly retain the original environment-variable flow:

export NEXSCOPE_AUTH_MODE="environment"
export NEXSCOPE_PROXY_BASE="https://api.nexscope.ai/"
export NEXSCOPE_API_KEY="<your_api_key>"

Never commit API keys or include them in prompts, logs, fixtures, or bug reports. There is no plaintext fallback when the system credential store is unavailable.

If the account has no active subscription, the authorization page does not activate the device and directs the user to subscribe. The plugin does not create or display a separate API key.

Other local commands are status, doctor, disconnect, and uninstall. Use disconnect --local-only only when the service is unreachable; the remote device seat then remains occupied until it is removed on the account page.

Commercial release build

Prepare reviewed offline wheels for both supported platforms, then run:

python3 scripts/build_release.py 1.1.0 \
  --wheel-dir /path/to/reviewed-wheels \
  --trusted-keys /path/to/reviewed-trusted-keys.json \
  --artifact-base-url https://github.com/lambbell/nexscope-amazon-intelligence/releases/download/v1.1.0
python3 scripts/self_check.py

The build emits the three frozen platform/architecture archives, an exact ManifestV1 metadata file, and a mode-0600 private payload-key file. It injects the reviewed public key set without storing production trust material in Git. Import the private file into the NexScope release service, then securely delete it.

Install locally

The commercial archive contains a thin installer, encrypted payload, reviewed offline dependencies, and installation instructions. Plaintext skills are installed only after the service grants a licensed payload key.

Publisher

ECOCREATE TECHNOLOGY PTE. LTD.
Support: service@nexscope.ai
Website · Privacy · Terms