Security fixes are provided for the latest tagged release and the main branch.

Do not open a public GitHub issue for security vulnerabilities.

Report vulnerabilities through GitHub Security Advisories. Include:

• affected version or commit SHA,
• deployment mode (http or stdio transport),
• relevant Docker Compose configuration with secrets redacted,
• steps to reproduce,
• observed impact,
• any known mitigation or workaround.

We will acknowledge valid reports as soon as possible, investigate privately, and publish a fix or advisory when appropriate.

By default, only web-mcp is published to the host (MCP at /mcp, REST at /api/v1/*); SearXNG, Crawl4AI, and Valkey stay on the Docker network. If you expose the server outside a trusted local environment, set API_TOKEN for REST endpoints and/or put the service behind a VPN, firewall rule, or authenticated reverse proxy.

web_extractor performs lightweight URL validation before calling extract providers. Crawl4AI domain/link filters are useful for crawl behavior, but they are not treated as this project's SSRF protection boundary.

Do not commit config/providers.yaml — it may contain API keys. Prefer ${ENV_VAR} references and keep .env out of version control. config/providers.yaml.example is safe to commit.

In scope:

• vulnerabilities in the web-mcp server (MCP and REST exposers),
• unsafe request handling in web_search or web_extractor,
• secret leakage in install, compose, logging, or docs,
• container configuration issues that expose internal services unexpectedly.

Out of scope:

• vulnerabilities in upstream SearXNG, Crawl4AI, Valkey, Docker, or Python dependencies unless this repo's configuration makes them exploitable in a new way,
• denial-of-service against a publicly exposed SearXNG or Crawl4AI instance that was intentionally exposed outside the compose network,
• reports that require access to a user's private deployment without authorization.