ClawHub

Free Web Browser

v0.1.0

Browse web through OpenClaw web_search and web_fetch with Ollama as search provider, no external API key required. Use when user asks to search web, verify c...

⭐ 1· 115·0 current·0 all-time
byPrantik Medhi@prantikmedhi

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for prantikmedhi/free-web-browser.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Free Web Browser" (prantikmedhi/free-web-browser) from ClawHub.
Skill page: https://clawhub.ai/prantikmedhi/free-web-browser
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install free-web-browser

ClawHub CLI

Package manager switcher

npx clawhub@latest install free-web-browser
Security Scan
Capability signals
Requires sensitive credentials
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
✓
Purpose & Capability
Name and description match the instructions: the skill exclusively relies on web_search and web_fetch (Ollama-backed) to discover and fetch pages. There are no unrelated env vars, binaries, or external services requested.
✓
Instruction Scope
SKILL.md only instructs the agent to rewrite queries, call web_search, pick URLs, and call web_fetch (extractMode markdown). It explicitly disallows shelling out to Ollama and does not ask the agent to read local files, secrets, or unrelated system state.
✓
Install Mechanism
No install spec and no code files — instruction-only skill. Nothing is written to disk and nothing is downloaded, which minimizes install risk.
✓
Credentials
Skill declares no required environment variables, credentials, or config paths. SKILL.md does not reference hidden env vars or secrets. Requested access is proportional to the stated browsing purpose.
✓
Persistence & Privilege
always is false and model invocation is allowed (the platform default). The skill does not request permanent presence or modify other skills or system settings.
Assessment
This skill is instruction-only and appears safe: it simply instructs the agent to use the platform's web_search and web_fetch tools (Ollama-backed) to find and fetch pages. Before installing, confirm your OpenClaw runtime actually exposes web_search/web_fetch and that Ollama is configured if you expect keyless operation. Remember that browsing live web content can surface inaccurate or sensitive information—avoid asking the agent to fetch pages containing your secrets or private internal URLs, and treat summaries of time-sensitive claims as potentially out of date.

Like a lobster shell, security has layers — review code before you run it.

latestvk977k3d2zs0e3x38r2rsqv0njx854mcj
115downloads
1stars
1versions
Updated 1w ago
v0.1.0
MIT-0
Prantik Medhi@prantikmedhi
latest
Download

Free Web Browser

Use web_search for discovery, then web_fetch for page content.

Quick workflow

  1. Rewrite vague request into concrete search query.
  2. Run web_search first.
  3. Pick 1 to 3 promising URLs.
  4. Run web_fetch on source pages you actually need.
  5. Cross-check important claims across more than one source.
  6. Summarize findings with links, caveats, and date sensitivity when relevant.

Query patterns

Prefer short, specific queries.

  • Current fact: site:docs.openclaw.ai web_search ollama
  • Product compare: ollama web search brave searxng comparison
  • Official docs first: site:github.com OR site:docs.openclaw.ai <topic>
  • Recent news: <topic> 2026

If first query weak, tighten scope instead of fetching random pages.

Tool choice

web_search

Use for:

  • finding sources
  • checking whether topic has changed recently
  • locating official docs, repos, blog posts, issues
  • getting top candidate URLs before deeper reading

web_fetch

Use for:

  • extracting readable page content
  • summarizing article, doc, or blog post
  • verifying exact wording from source page
  • pulling details after search already found target URL

Prefer extractMode: "markdown" unless plain text is better for logs or machine-like parsing.

Quality rules

  • Prefer official docs, project repos, vendor docs, standards bodies, and primary sources.
  • Do not trust snippets alone for important claims. Fetch source page.
  • For time-sensitive topics, say info may change and mention recency.
  • When sources disagree, say so plainly.
  • Avoid unnecessary fetch spam. Search once, fetch only best pages.

Ollama-specific guidance

This skill assumes web_search is configured to use Ollama-backed search, so no separate search API key is needed.

  • Use normal web_search tool. Do not shell out to Ollama manually unless debugging config.
  • If search fails, suspect provider setup before blaming query.
  • If Ollama search returns thin snippets, fetch source pages for substance.
  • If current runtime does not expose web_search, this skill cannot browse live web.

Failure handling

  • No results: broaden query once, then try official-site search.
  • Low-quality results: add site filters or product/version terms.
  • Fetch blocked or empty: choose alternate source from search results.
  • Ambiguous request: ask one clarifying question only after one good-faith search attempt.

Output pattern

Default answer shape:

  • direct answer first
  • 2 to 5 bullets with evidence
  • source links
  • brief uncertainty note if needed

Read next

For troubleshooting and example prompts, read references/ollama-web-browser-notes.md.

Comments

Loading comments...