Security

What the source does

• FilesystemNone observedNo reads or writes in the source
• NetworkNone observedNo outbound calls in the source
• ShellNone observedNo commands run in the source
• DependenciesPinnedNothing floats
• SecretsNone foundNo keys in source

Five audit layers

Gates applied: no behavioural pass.

HTTPoison.get(url)  # User passes http://169.254.169.254/metadata

HTTPoison.get(url)  # User passes http://169.254.169.254/metadata

Audit 0.4.1 · source sha 492f17636999 · 18 Oct 2026Report an issue or request a re-scan

• Publisheroliver-kriska GitHub userOwner of the repository this skill is read from
• Source folderoliver-kriska/claude-elixir-phoenix/plugins/elixir-phoenix/skills/securityThe exact folder we audited
• LicenceMITAs the repository states it
• Commit pinned for this verdict492f17636999A new audit runs when the source changes
• Repository statusNot archivedLast push 2 days ago
• Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

Its description: Build and harden Phoenix auth and security — OAuth login, password hashing, sessions, RBAC, rate limiting, CSRF, XSS, SQL injection, secrets. Use when adding login flows or permissions, or handling user input.

---
name: security
description: "Build and harden Phoenix auth and security — OAuth login, password hashing, sessions, RBAC, rate limiting, CSRF, XSS, SQL injection, secrets. Use when adding login flows or permissions, or handling user input."
effort: medium
user-invocable: false
paths:
  - "**/*auth*.ex"
  - "**/*session*.ex"
  - "**/*password*.ex"
---

# Elixir/Phoenix Security Reference

> **Ash projects**: `AshAuthentication` has its own strategy/token patterns — use the `ash-framework` skill. CSRF, XSS, and secret management patterns below still apply.

Quick reference for security patterns in Elixir/Phoenix.

## Iron Laws — Never Violate These

1. **VALIDATE AT BOUNDARIES** — Never trust client input. All data through changesets
2. **NEVER INTERPOLATE USER INPUT** — Use Ecto's `^` operator, never string interpolation
3. **NO String.to_atom WITH USER INPUT** — Atom exhaustion DoS. Use `to_existing_atom/1`
4. **AUTHORIZE EVERYWHERE** — Check in contexts AND re-validate in LiveView events
5. **ESCAPE BY DEFAULT** — Never use `raw/1` with untrusted content
6. **SECRETS NEVER IN CODE** — All secrets in `runtime.exs` from env vars
7. **LIVEVIEW EVENT PARAMS ARE UNTRUSTED** — Users can alter forms, hooks, and every `phx-value-*` in DevTools. Validate and authorize against server-side state before acting

## Quick Patterns

### Timing-Safe Authentication

```elixir
def authenticate(email, password) do
  user = Repo.get_by(User, email: email)

  cond do
    user && Argon2.verify_pass(password, user.hashed_password) ->
      {:ok, user}
    user ->
      {:error, :invalid_credentials}
    true ->
      Argon2.no_user_verify()  # Timing attack prevention
      {:error, :invalid_credentials}
  end
end
```

### LiveView Authorization (CRITICAL)

```elixir
# `id` is client input even when it came from phx-value-id.
# RE-AUTHORIZE IN EVERY EVENT HANDLER
def handle_event("delete", %{"id" => id}, socket) do
  post = Blog.get_post!(id)

  # Don't trust that mount authorized this action!
  with :ok <- Bodyguard.permit(Blog, :delete_post, socket.assigns.current_user, post) do
    Blog.delete_post(post)
    {:noreply, stream_delete(socket, :posts, post)}
  else
    _ -> {:noreply, put_flash(socket, :error, "Unauthorized")}
  end
end
```

Rendered LiveView events can expose IDs in HTML and websocket payloads. That is
not automatically a vulnerability: treat IDs as public identifiers, never as
proof of access. Use opaque references only when the identifier itself must not
be disclosed, and still perform server-side authorization.

### SQL Injection Prevention

```elixir
# ✅ SAFE: Parameterized queries
from(u in User, where: u.name == ^user_input)

# ❌ VULNERABLE: String interpolation
from(u in User, where: fragment("name = '#{user_input}'"))
```

## Quick Decisions

### What to validate?

- **All user input** → Ecto changesets
- **File uploads** → Extension + magic bytes + size
- **Paths** → `Path.safe_relative/2` for traversal
- **Atoms** → `String.to_existing_atom/1` only

### What to escape?

- **HTML output** → Auto-escaped by default (`<%= %>`)
- **User HTML** → HtmlSanitizeEx with scrubber
- **Never** → `raw/1` with untrusted content

## Anti-patterns

| Wrong | Right |
|-------|-------|
| `"SELECT * FROM users WHERE name = '#{name}'"` | `from(u in User, where: u.name == ^name)` |
| `String.to_atom(user_input)` | `String.to_existing_atom(user_input)` |
| `<%= raw @user_comment %>` | `<%= @user_comment %>` |
| Hardcoded secrets in config | `runtime.exs` from env vars |
| Auth only in mount | Re-auth in every `handle_event` |
| Trusting `phx-value-*` or hidden IDs | Load server-side state and authorize it |

## References

For detailed patterns, see:

- `${CLAUDE_SKILL_DIR}/references/authentication.md` - phx.gen.auth, MFA, sessions
- `${CLAUDE_SKILL_DIR}/references/authorization.md` - Bodyguard, scopes, LiveView auth
- `${CLAUDE_SKILL_DIR}/references/input-validation.md` - Changesets, file uploads, paths
- `${CLAUDE_SKILL_DIR}/references/security-headers.md` - CSP, CSRF, rate limiting, headers
- `${CLAUDE_SKILL_DIR}/references/oauth-linking.md` - OAuth account linking, token management
- `${CLAUDE_SKILL_DIR}/references/rate-limiting.md` - Composite key strategies, Hammer patterns
- `${CLAUDE_SKILL_DIR}/references/advanced-patterns.md` - SSRF prevention, secrets management, supply chain

Excerpt, first 4,354 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

git clone https://github.com/oliver-kriska/claude-elixir-phoenix
git -C claude-elixir-phoenix checkout 492f17636999

Works in

1. 2026-10-18
   ✓SAFE B 89 / 100

   first audit · source 492f17636999

Questions about Security

• github.com/oliver-kriska/claude-elixir-phoenix/tree/main/plugins — the audited source at commit 492f17636999 and every finding aboveObserved · 18 Oct 2026

Spotted something wrong? Report a correction · Methodology