agent-skills (ogulcancelik)
Two small skills: web search through your own browser with no paid API, and a preflight risk check before a meaningful change.
- 81 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The daemon launches a headless browser and reaches the web from your address.
- The skill opens arbitrary pages and their content enters the agent context.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/agent-skills-ogulcancelikDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The daemon launches a headless browser and reaches the web from your address.
- The skill opens arbitrary pages and their content enters the agent context.
This is third-party code. Review the repository files before installing.
What it does
The set holds two skills. `web-search` gives an agent free local web search: a background daemon keeps a headless browser, searches Google with a DuckDuckGo fallback, opens selected results, renders JavaScript pages and returns clean Markdown. A CLI exists: a query, a result count with `-n`, result selection through `--from` and `--fetch`, and direct opening of an address with `--url`. `preflight` sets up a risk-proportional process around a meaningful implementation: before code it challenges scope, afterwards it verifies the change against a locked contract. The author marks it experimental, so its shape may change a lot.
Who it is for. Developers who want web search for a coding agent without search API keys and a check on task size before edits begin.
Good fit when
- The agent needs fresh documentation and you do not want a paid search API.
- A page renders with JavaScript and a plain HTTP request returns nothing.
- Before migrating to another library you ask `preflight` to challenge the scope.
- After an edit you need to check the result against an agreed contract rather than the agent memory.
Not a fit when
- No Chromium-family browser is installed.
- Search runs in a closed network or over corporate documents where Google is unavailable.
- You need a stable process: `preflight` is still experimental.
Example request
Search the web for what changed in the latest Vite version, open the first three results and give a Markdown digest.Limitations
For `web-search` you need Node.js 22.23.4, 24.22.1 or 26.12.1 and newer (a Brotli fix) and a Chromium-family browser: Chrome, Brave, Edge or Chromium. The agent installs dependencies with Bun. The daemon listens on local ports, 9377 by default and 9225 for DevTools. The skill recognises bot-protection pages and does not guarantee opening sites that reject simple requests. Queries go to Google or DuckDuckGo from your address. The author considers `preflight` experimental.
How to disable. Delete the `web-search` and `preflight` skill folders from the agent skills directory and stop the local search daemon.
FAQ
How do I install the skills?
`npx skills add ogulcancelik/agent-skills` installs the whole set. Pick one with `--skill web-search` or `--skill preflight`.
Which variables configure search?
`WEB_SEARCH_BROWSER_BIN` sets the browser path, `WEB_SEARCH_USER_AGENT` overrides the user agent, `WEB_SEARCH_DAEMON_PORT` and `WEB_SEARCH_CDP_PORT` move ports, and `WEB_SEARCH_DEBUG_DUMP` saves artifacts after failures.
Which systems does web-search run on?
Linux, macOS and Windows. The CLI finds Chrome, Brave, Edge or Chromium in common install locations.
What does preflight do?
Before code it challenges the scope, afterwards it verifies the result against a locked contract. Depth depends on risk.
Which search engines does web-search use?
Google is the main search with DuckDuckGo as a fallback. Result pages open in a local browser and are reduced to Markdown.
Discussion
No questions or reviews yet.
Related
SkillsA web research skill: SearXNG searches, Camofox reads pages and CloakBrowser opens sites that check their visitors
Three skills and four utilities that let an agent search the web, read pages, pull video subtitles and process media
A skill connects an agent to your Chrome through an extension and CLI for reading pages, managing tabs and verifying actions.
Three search skills: four-source search scored by intent, deep thread tracing and clean Markdown extraction from pages.
A skill for web search, X search and page fetching for models without web access: no signup or key, the answer is JSON with source links
Web Access
web-access
A web access skill that picks between search, page fetch and driving your own browser over CDP
README in short
A set of small skills in the Agent Skills format. Each installs separately, and `web-search` has its own README with settings.