LLM-1285: Web fetch tool

[pmateusz]

Summary

Adds a web_fetch tool to kimchi-code as an in-repo extension (extensions/web-fetch/). The tool fetches web pages — including JavaScript-rendered SPAs — and converts HTML to clean markdown for the coding agent to consume.

What it does

• Fetches any URL using Playwright (headless Chromium) as the primary fetcher, with automatic fallback to native fetch() when Playwright/Chromium is not installed
• Converts HTML to markdown (default), plain text, or raw HTML via Turndown
• Browser pool — single Chromium instance with lazy init, tab reuse (~100ms vs 1-2s cold start), idle timeout (60s), and crash recovery
• Session-scoped cache with 15-min TTL and LRU eviction (100 entries) — enables drill-down workflows without re-fetching
• SSRF protection — blocks private IPs (10.x, 172.16-31.x, 192.168.x, 127.x, 169.254.x, ::1, fc00::/7), localhost, and cloud metadata endpoints
• Response size limit (5MB), output truncation, configurable timeout (default 30s, max 120s)
• Strips script/style/meta tags before conversion
• Follows redirects automatically, reports final URL in metadata

Tool parameters

Parameter	Type	Default	Description
url	string (required)	—	URL to fetch (http:// or https://)
format	markdown | text | html	markdown	Output format
timeout	number	30	Timeout in seconds (max 120)

[wiz-3b6633f35c]

Wiz Scan Summary

Scanner	Findings
Vulnerabilities	-
Sensitive Data	-
Secrets	-
IaC Misconfigurations	-
SAST Findings	1  2
Software Management Findings	-
Total	1  2

View scan details in Wiz

To detect these findings earlier in the dev lifecycle, try using Wiz Code VS Code Extension.

[wiz-3b6633f35c]

 SAST Finding

Server-Side Request Forgery (SSRF) via User-Controlled URLs in Node.js HTTP Client Libraries (CWE-918)
Analyzed by AI as

More Details

This rule detects instances where user-controlled URLs are passed directly to HTTP client libraries, leading to a Server-Side Request Forgery (SSRF) vulnerability. SSRF allows an attacker to abuse server functionality to make requests to internal systems or services that are not directly exposed to the internet. This can result in unauthorized access to sensitive data, internal network reconnaissance, and potential pivoting into the internal network.

Attribute	Value
Impact
Likelihood

AI Analysis

The code snippet shows a fetch() call with a url parameter, but there's insufficient context to determine if this URL comes from user input or is properly validated. The snippet doesn't show the source of the url variable or any validation logic. Without seeing how url is populated or if there are security controls elsewhere in the function, it's impossible to definitively classify this as exploitable or safe.

Remediation

Server-Side Request Forgery (SSRF) is a vulnerability that allows an attacker to force the server to make unintended requests to other systems or internal resources. This can lead to unauthorized access, data theft, and potentially further compromise of the internal network.

To mitigate SSRF, user input should never be directly passed to HTTP client libraries. Instead, implement strict input validation and whitelisting to ensure that only trusted URLs are allowed. This can be done by maintaining a list of approved domains or URLs, and rejecting any user input that does not match this whitelist.

Code examples:

// VULNERABLE CODE - User input is directly passed to the HTTP client library
const axios = require('axios');
const userUrl = req.query.url;

axios.get(userUrl)
.then(response => {
// Handle response
})
.catch(error => {
// Handle error
});

// SECURE CODE - User input is validated against a whitelist before making the request
const axios = require('axios');
const userUrl = req.query.url;
const allowedUrls = ['https://example.com', 'https://example.org'];

if (allowedUrls.includes(userUrl)) {
axios.get(userUrl)
.then(response => {
// Handle response
})
.catch(error => {
// Handle error
});
} else {
// Reject the request with an appropriate error message
}

Additional recommendations:

• Follow the principle of least privilege and only grant the minimum required permissions to the server.
• Implement strict input validation and sanitization for all user input, not just URLs.
• Keep software up-to-date and apply security patches promptly.
• Adhere to the OWASP Server-Side Request Forgery Prevention Cheat Sheet and other relevant security standards.
• Consider using a secure web gateway or web application firewall (WAF) to provide an additional layer of protection.

Rule ID: WS-I007-JAVASCRIPT-00019

To ignore this finding as an exception, reply to this conversation with #wiz_ignore reason

If you'd like to ignore this finding in all future scans, add an exception in the .wiz file (learn more) or create an Ignore Rule (learn more).

To get more details on how to remediate this issue using AI, reply to this conversation with #wiz remediate