EveryInfra Analysis Skills

English · 简体中文 · Español

Free to download. Use your own AI tool and model.

EveryInfra Analysis Skills is version 0.2.0, licensed under Apache-2.0. The package is free to download and install. It ships four analysis Skills. It does not call a model, does not bundle a model provider, and has no API key or runtime dependency. The customer's compatible AI tool supplies the model and the model quota.

A separate EveryInfra data-collection API or MCP call may cost money. Free Skill installation is not free collection.

Skills

Slug	What it analyzes
ecommerce-product-analysis	Product candidates from user-supplied CSV/JSON or results already fetched.
social-content-performance	Social content for a stated sample and time window, from supplied files or results already fetched.
content-analysis-tagging	Records labeled with a user-provided or agreed taxonomy.
review-insights	Pain points, praise, and rating patterns in product, app, place, or company reviews, from supplied files or results already fetched.

Each Skill analyzes supplied files or already-fetched results. Installing a Skill does not collect data.

Use with EveryInfra data

Each Skill also works on results you already fetched from EveryInfra. Examples of matching EveryData actions (the live catalog lists current actions and fields):

Skill	Example EveryData actions
ecommerce-product-analysis	amazon/search, walmart/search, taobao/search, jd/search, shopee/search
social-content-performance	tiktok/search, tiktok/user_videos, youtube/search, xiaohongshu/search, bilibili/search
content-analysis-tagging	Text records such as xiaohongshu/search, reddit/search, zhihu/search, or comments results
review-insights	amazon/reviews, walmart/reviews, taobao/reviews, jd/reviews, app_store/reviews, google_maps_reviews/reviews

Live catalog: GET https://api.everyinfra.com/api/v1/social/catalog.

Quick start

1. Choose the Skill for the task.
2. Download the repository ZIP or clone https://github.com/everyinfra/analysis-skills.git. Install the selected Skill folder to the compatible host's documented Skill location. This README names ~/.agents/skills/ and ~/.claude/skills/.
3. Configure or select the customer's own model in the AI host's own settings. The package does not configure the host and does not provide a model key.
4. Open the matching synthetic fixture. Give the Skill the fixture's source_response and task. The fixture includes expected_output; compare the returned structure, not exact model wording. This step makes no EveryInfra data-collection call.
5. For new data collection, see the live API/MCP documentation. Collection is separate and may be billable.

Download

Obtain the repository source ZIP from the GitHub repository page:

1. Open https://github.com/everyinfra/analysis-skills.
2. Use GitHub Code → Download ZIP.
3. Unpack the ZIP.

Or clone https://github.com/everyinfra/analysis-skills.git.

Install

Host	Install Directory
Codex CLI / IDE	~/.agents/skills/
Claude Code	~/.claude/skills/

Copy one or more of these exact Skill folders into ~/.agents/skills/ or ~/.claude/skills/:

• skills/ecommerce-product-analysis
• skills/social-content-performance
• skills/content-analysis-tagging
• skills/review-insights

Each Skill folder is independently installable. Compatible hosts can use their own documented Skill folder.

Codex CLI / IDE: $ecommerce-product-analysis · $social-content-performance · $content-analysis-tagging · $review-insights

Claude Code: /ecommerce-product-analysis · /social-content-performance · /content-analysis-tagging · /review-insights

For a synthetic example, give the model only source_response and task; retain expected_output outside the model input to compare afterward.

Build skills — Codex · Extend Claude with skills

Synthetic fixtures

The examples/ directory contains four synthetic fixtures, one per Skill. They are synthetic:

• examples/ecommerce-product-analysis.synthetic.json
• examples/social-content-performance.synthetic.json
• examples/content-analysis-tagging.synthetic.json
• examples/review-insights.synthetic.json

Open the matching fixture and give the Skill the fixture's source_response and task. Each fixture includes expected_output. Compare the returned structure, not exact model wording. Using a fixture makes no EveryInfra data-collection call.

Collection and public API

Analysis in these Skills is free of Skill charges. Collection through EveryInfra API or MCP is a separate step and may be billable. Catalog or tool discovery does not grant permission to make a paid data call.

Public API references:

• https://api.everyinfra.com/docs
• https://api.everyinfra.com/openapi.json

Maintenance and feedback

Maintainer: EveryInfra (the organization).

Update triggers: public API contract or schema changes, a verified correction to a Skill or example, or an accepted user request. At each package version update, review open feedback before changing the manifest version and changelog. This is an event-driven manual cadence.

Feedback channel: public GitHub Issues at https://github.com/everyinfra/analysis-skills/issues. Do not post API keys, credentials, private customer records, or other confidential data.

License

Apache-2.0. See LICENSE.