← All skills
openweb
80 / 100 · Low risk
Fine to try. Some of what it can do is expected: a marketing skill usually talks to online services.
By openweb-org/openweb (GitHub) · marketing · 77 stars · scanned at commit a387b50
Typed JSON access to 90+ real websites — Google, Amazon, Reddit, YouTube, GitHub, Instagram, Bloomberg, Zillow, and more — across search, shopping, travel, finance, social, news, and dev tools. Use to read, search, post, comment, message, or otherwise interact with these sites (prices, products, articles, stock quotes, flights, posts, comments, DMs, etc.). Run `openweb sites` for the full list.
Works with any agent that supports SKILL.md skills.
What this skill can do
Runs commands on your computer
The author does not mention this.
Found in 4 places.
Show 1 more
Connects to the internet
The author does not mention this. Expected for a marketing skill.
Found in 37 places.
Show 17 more
- skills/openweb/add-site/capture.md, line 146
await page.evaluate(() => fetch('/api/v1/feed?limit=20', { - skills/openweb/add-site/capture.md, line 220
const r = await fetch(args.url, { - skills/openweb/add-site/curate-runtime.md, line 57
- **`page`** — Executes `fetch()` inside the browser tab. Required when Akamai, - skills/openweb/add-site/probe.md, line 109
const r = await fetch(url, { credentials: 'include' }) - skills/openweb/knowledge/adapter-recipes.md, line 86
- A simple `fetch()` with a User-Agent header returns the full page - skills/openweb/knowledge/adapter-recipes.md, line 390
return fetch(url, { - skills/openweb/knowledge/adapter-recipes.md, line 395
'X-Requested-With': 'XMLHttpRequest', - skills/openweb/knowledge/bot-detection.md, line 25
- **APIRequestContext also fails on write paths:** Even with `page` transport and a logged-in session, `page.request.fetch()` (Playwright APIRequestContext) can - skills/openweb/knowledge/bot-detection.md, line 33
- **API call blocking:** On aggressive PX sites (e.g. Zillow), both `page.evaluate(fetch())` AND `page.request.fetch()` are blocked — PX validates at network/co - skills/openweb/knowledge/bot-detection.md, line 69
- **Impact:** `page.evaluate(fetch(...))` inherits the signing because it runs through the patched `fetch`. Node `fetch` does not have the patch and will fail. - skills/openweb/knowledge/bot-detection.md, line 71
- **Verification:** Must verify via probe -- check `fetch.toString().length` in browser devtools before deciding transport. The monkey-patch may also exist on ` - skills/openweb/knowledge/bot-detection.md, line 90
| Custom required headers | e.g. some social media, image-sharing sites | Requires `x-requested-with: XMLHttpRequest` or site-specific app ID headers -- 400/403 - skills/openweb/knowledge/bot-detection.md, line 108
+- Does page.evaluate(fetch(...)) work? - skills/openweb/knowledge/bot-detection.md, line 110
+- No -> Does page.request.fetch() work? - skills/openweb/knowledge/bot-detection.md, line 111
+- Yes -> adapter with page.request.fetch (e.g. Costco read ops; cart writes switch to page.evaluate(fetch()) due to Akamai header-bundle fingerprinting) - skills/openweb/knowledge/bot-detection.md, line 120
but also client-side sensor data attached to each request. `page.evaluate(fetch(...))` - skills/openweb/knowledge/graphql.md, line 13
- **Hash rotation:** Some sites rotate persisted query hashes on every deploy. For these, hardcoding hashes is fragile. Use an L3 adapter that extracts hashes a
and 17 more places
Changes files outside its own folder
The author does not mention this.
Found in 1 place.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'openweb' ] || [ -L ~/.agents/skills/'openweb' ]; then echo 'A skill folder named openweb already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/openweb-org/openweb' "$d" && git -C "$d" checkout --quiet 'a387b50c829d871839a613732e1b97bfa1946124' -- 'skills/openweb' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/openweb' ~/.agents/skills/'openweb' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'openweb' ~/.claude/skills/'openweb' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for openweb-org/openweb shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/openweb-org/openweb/)
Score history
| Date | Version | Score |
|---|
| 2026-11-07 | a387b50 | 80 |