# Security Policy

Security guidelines and vulnerability reporting for OwnCord.

## Reporting Vulnerabilities

Report privately via GitHub Security Advisories:
[github.com/J3vb/OwnCord/security/advisories/new](https://github.com/J3vb/OwnCord/security/advisories/new).

**Do NOT open public issues for security bugs.**

The repository-root [SECURITY.md](../SECURITY.md) is the canonical reporting
policy — what to include and the response timeline (initial response within
7 days) live there, so the two files cannot disagree.

## What stays private, and for how long

This applies to weaknesses in the repository's own automation and settings —
workflow authorization, credential scope, release gating — as much as to bugs in
the server or client. Planning documents cite this section as the rule; it is
written here so the citation points at something.

- Public artifacts — commits, issues, pull request descriptions, changelogs —
  carry an **opaque identifier, the affected property, safe acceptance criteria,
  and a status**. Nothing more.
- Reproduction steps, source-to-sink traces, exploit conditions, and the state a
  fix replaced stay in the private advisory. A commit that fixes a weakness
  describes the control it adds, not the gap it closes.
- Every private finding has exactly one public owner, so nothing is tracked only
  in private and nothing is silently dropped.
- Release notes may describe repaired impact after coordinated remediation,
  without the detail needed to reproduce it.

This repository is public. A commit message is a disclosure channel.

## Trust model

What the server operator can and cannot read, what is end-to-end encrypted,
how transport and at-rest data are protected, and what beta does not claim are
stated in one place: [trust-model.md](trust-model.md). Every claim there cites
the code line or test that makes it true.

## Two-Factor Authentication

OwnCord supports TOTP-based 2FA:

- Users enroll via Settings > Account (QR code + backup codes)
- Admins can enforce server-wide 2FA via the `require_2fa` setting in the admin panel
- `require_2fa` requires all users to have 2FA enabled and registration to be closed
- Login flow returns `requires_2fa: true` with a `partial_token` (10-min TTL, 5-attempt limit)
- Auth challenges are rate-limited to 10 req/min per IP
- TOTP code verification uses constant-time comparison (`subtle.ConstantTimeCompare`) to prevent timing side-channel attacks

## Account Deletion

Users can delete their own account via `DELETE /api/v1/auth/account` with password confirmation. The last admin account cannot be deleted. After 3 failed password attempts, the endpoint locks out for 15 minutes.

## Audit Logging

Security-relevant actions are recorded in the `audit_log` table with actor, action, target, and detail:

- **Auth:** `user_register`, `user_login`, `user_logout`, `login_blocked_banned`, `account_deleted`, `password_change`, `session_revoke`
- **2FA:** `totp_enabled`, `totp_verified`, `totp_disabled`
- **Admin:** `role_change`, `role_create`, `role_update`, `role_delete`, `role_reorder`, `user_ban`, `user_unban`, `force_logout`, `setting_change`, `server_setup`, `api_token_create`, `api_token_revoke`, `config_write`, `invite_create`, `invite_revoke`, `plugin_install`, `plugin_uninstall`
- **Content:** `channel_create`, `channel_update`, `channel_delete`, `channel_perms_update`, `channel_perms_clear`, `channel_user_perms_update`, `channel_user_perms_clear`, `message_delete`, `message_purge`, `emoji_create`, `emoji_delete`
- **Profile:** `profile_update`, `identity_key_update`
- **Ops:** `backup_create`, `backup_delete`, `backup_restore`, `ws_connect`

Note: `backup_restore` is written synchronously to the live database _before_
the pre-restore safety copy is taken, so the row survives inside the
`pre_restore_*.db` backup. The restored database itself will not contain it —
the restore replaces the database file wholesale.

## Client Security Hardening

The Tauri desktop client implements the following security measures:

### Credential Storage

- Credentials are stored in the OS keyring (Windows Credential Manager / macOS Keychain / Secret Service) via the `keyring` crate, with every write read back and verified; if no keyring is available they fall back to an encrypted file (Windows DPAPI with `CRYPTPROTECT_UI_FORBIDDEN`, ChaCha20-Poly1305 elsewhere) — see [credential-storage.md](credential-storage.md)
- Plaintext passwords are **never** returned to the frontend over IPC — only tokens are accessible from JavaScript
- Auto-login uses stored tokens for reconnection, not passwords

### Tauri Capabilities (Least Privilege)

- Filesystem write access is scoped to `$APPDATA/**` and `$APPLOG/**` only
- DevTools command is gated behind the `devtools` feature flag (excluded from release builds)
- HTTP fetch is restricted to `https://` origins plus `http://127.0.0.1:*` (the Rust TOFU proxy's loopback tunnel), and **denies** `https://localhost[:*]` and `https://127.0.0.1[:*]` — no legitimate flow reaches loopback over https, so the deny list keeps the renderer from probing other local services
- `http:allow-fetch` is the **only** URL-scoped HTTP identifier. `tauri-plugin-http` validates the URL exactly once, in the `fetch` command; `fetch_send` and `fetch_read_body` operate on an already-validated `ResourceId` and never consult a scope, so `allow`/`deny` blocks on those identifiers are inert and were removed rather than left in place advertising a control that does not exist
- The `https://*` wildcard cannot be removed today: link previews (`embeds.ts`) fetch arbitrary user-posted URLs by design, and Tauri scopes per _command_, not per JS caller. Bounded in TypeScript by `isPrivateHost`/`isBlockedForPreview`, a 5 s timeout and a 50 KB body cap; the response is regex-scraped for `og:` tags and never executed
- Regression-guarded by `tests/unit/capabilities-scope.test.ts`; rationale and the follow-up that would remove the wildcard are in [docs/plans/tauri-capability-narrowing.md](plans/tauri-capability-narrowing.md)

### TLS and Certificate Pinning (TOFU)

- Self-signed certificates are supported via Trust-On-First-Use (TOFU) pinning
- The WebSocket proxy (`ws_proxy`) pins the server certificate fingerprint on first connection
- The LiveKit proxy (`livekit_proxy`) reuses the pinned fingerprint from the WS proxy
- Certificate mismatch triggers a modal requiring user acknowledgment
- Update downloads validate `server_url` uses `https://` and rejects URLs with userinfo

### Input Validation

- IPC commands validate host format, string lengths, and character allowlists
- PTT virtual key codes are validated to the Win32 range (1–254)
- LiveKit proxy `remote_host` is validated against CRLF injection
- API client validates host format before constructing URLs
- File uploads enforce a MIME type allowlist (images, video, audio, PDF, text)
- Error messages from server responses are capped at 200 characters
- Notification titles are sanitized (control chars stripped, length capped)

### XSS Prevention

- All user-generated content is rendered via `textContent`/`setText` — never `innerHTML`
- The single `innerHTML` usage (SVG icons) operates on compile-time constants with a runtime guard
- URLs are validated via `isSafeUrl` (rejects `javascript:`, `data:`, `vbscript:`)
- YouTube embeds use `sandbox` attribute on iframes
- `image/svg+xml` is excluded from safe MIME types for data URIs
- GIF media URLs are validated against the trusted Klipy CDN origins
- Linkified URLs strip trailing punctuation to prevent misleading destinations

### Search and Rate Limiting

- Client-side search requests are rate-limited (500ms minimum interval + 300ms debounce)

## Known Limitations

- Server auto-updates depend on a dedicated pinned minisign/Ed25519 server release key in [Server/updater/server_update_public_key.txt](../Server/updater/server_update_public_key.txt) and a signed release manifest that binds the shipped binary hash to the release version; Windows Authenticode/SmartScreen code signing is still separate work
- CSP `connect-src` allows `https:` to any host (necessary for self-hosted server URLs not known at build time). Because of this, narrowing the Tauri `http:allow-fetch` scope alone would not bound exfiltration from a compromised renderer — the webview's own `fetch` reaches the same hosts without going through the plugin. Closing that requires narrowing `connect-src` and moving the link-preview fetch into Rust in the same change

## Security Hardening Checklist for Operators

- [ ] Enable TLS (self-signed is the default; custom certs recommended for production)
- [ ] Keep invite-only registration enabled (default)
- [ ] Set a strong admin password
- [ ] Configure rate limits (defaults are sensible but review for your use case)
- [ ] Run regular backups via the admin panel
- [ ] Keep the server updated (admin panel shows available updates)
- [ ] Firewall: only expose port 8443 (HTTPS); for voice/video also 7880-7881/TCP and 50000-60000/UDP (LiveKit signaling + media — see [deployment.md](deployment.md)); port 80 only when using ACME
- [ ] Enable server-wide 2FA requirement once all users have enrolled
- [ ] Set `admin_allowed_cidrs` to restrict admin panel access to trusted networks