Skip to content

Security: TraceHood/robinhood-explorer

Security

docs/SECURITY.md

Security

Product posture

Read-only MVP. No wallet connect for signing. No private keys, seed phrases, or transaction construction helpers that submit trades.

Threats we design for

  • Impersonation of tickers and look-alike addresses: always show checksummed address on detail views; warn on homoglyph/near-duplicate symbols.
  • Malicious token metadata (SVG, HTML, javascript: URLs, oversized images): sanitize and proxy.
  • SSRF via image URLs: proxy with allowlist schemes, block link-local/metadata IPs, timeout, MIME sniff, size cap.
  • Secret leakage: env-only secrets, no NEXT_PUBLIC_ for keys, redact logs.
  • User deanonymization: analyze public onchain behavior only; no offchain identity stitching.
  • Cluster overclaim: inferences are labeled; confirmed requires authoritative public evidence.

Web controls (Phase 1)

  • CSP on Next.js responses (default-src 'self', restricted img/connect).
  • Global search and graph queries are rate-limited at the API (demo in-process; production Redis).
  • Demo mode never silently presents fixtures as live.

Later

  • Image proxy service.
  • Audit log for registry, labels, methodology, and rules.
  • Public label challenge/correction workflow.
  • Webhook signing secrets, retry, dead-letter.
  • Managed secret store in production.

Disclosures

Terms, Privacy, Methodology, and Independent Project Disclaimer are first-class routes.

There aren't any published security advisories