Paste a secret, get a link, send it. The first person to open it and press
Reveal sees the secret; the link dies at that moment. The recipient needs a
browser and nothing else — no account, no client, no installed tooling.
The server cannot read what it stores. AES-256-GCM happens in the browser and
the key lives in the URL fragment, which browsers never transmit, so hushd
holds ciphertext and no key material. That is a property of where the key sits
rather than a promise about our conduct, which is why there is deliberately no
endpoint accepting a plaintext secret and no server-side-encryption fallback:
two guarantees behind one URL would be worse than one honest guarantee.
Three decisions carry the design:
* GET /s/{id} touches NO storage, not even to check existence. Slack, Teams,
WhatsApp, iMessage and Outlook Safe Links all fetch a URL before a human
sees it, so destroying on GET would destroy most secrets in transit and the
recipient's "already used" would be indistinguishable from interception.
Only POST /reveal consumes. Bot user-agent detection is an arms race;
removing the side effect from GET is not. Pinned by
TestGettingTheRevealPageNeverConsumesTheSecret.
* Destruction is one Redis GETDEL, which is atomic. GET-then-DEL has a window
where two simultaneous readers both win, and for a one-time secret that
window is the product. The store contract demands atomicity and the same
concurrency test runs against both implementations.
* Missing, already-revealed, expired and evicted are ONE indistinguishable
410. Separating them would confirm to a prober that a given link was real.
The secret id IS the capability, so secret.ID is a struct whose every
accidental path — %v, %s, String(), slog, json.Marshal — emits a redacted
handle or refuses, and the raw value needs an explicit Value(). The first
version tried to prevent leaks by implementing no String() at all; its own test
caught that Go's fmt prints unexported fields anyway, so forbidding the method
had removed the control rather than the leak.
Operationally: structured JSON on stdout in the fleet's wire format, which
Vector already collects with no annotation; six hush_* metrics on the chassis
registry with no id, IP or path in any label; five alert rules wired into
vmalert. The public Ingress enumerates /, /s/ and /api/ so /metrics, /healthz
and /readyz share the port but are unreachable from the internet — no
basic-auth middleware to maintain and get wrong.
Dependencies are vendored because go-chassis is private: the Woodpecker test
step and the in-cluster Kaniko build both run -mod=vendor with GOPROXY=off and
hold no git credential.
cmd/hush-mcp is a stdio MCP server doing the same client-side crypto locally,
so using hush from an agent preserves the same guarantee as using it from a
browser.
105 lines
3.5 KiB
Go
105 lines
3.5 KiB
Go
package chassis
|
|
|
|
import (
|
|
"context"
|
|
"crypto/subtle"
|
|
"net/http"
|
|
"strings"
|
|
)
|
|
|
|
// Identity is the authenticated principal attached to the request context by an
|
|
// auth Middleware. Subject is the stable principal id; OrgID is the tenant the
|
|
// principal is acting in (the multi-tenant scoping key — every tenant row
|
|
// filters by it); Scopes are the granted permissions; Claims carries any extra
|
|
// verified attributes without coupling chassis to a domain type.
|
|
type Identity struct {
|
|
Subject string
|
|
OrgID string
|
|
Scopes []string
|
|
Claims map[string]string
|
|
}
|
|
|
|
// HasScope reports whether the identity holds the required scope. A held scope
|
|
// ending in ":*" dominates any required scope sharing its prefix, so "admin:*"
|
|
// satisfies "admin:read". An exact match always passes.
|
|
func (id *Identity) HasScope(required string) bool {
|
|
if id == nil {
|
|
return false
|
|
}
|
|
for _, s := range id.Scopes {
|
|
if s == required {
|
|
return true
|
|
}
|
|
if prefix, ok := strings.CutSuffix(s, ":*"); ok {
|
|
if reqPrefix, _, found := strings.Cut(required, ":"); found && reqPrefix == prefix {
|
|
return true
|
|
}
|
|
}
|
|
}
|
|
return false
|
|
}
|
|
|
|
// Authenticator verifies a request and returns its Identity, or an *Error
|
|
// (Unauthorized/Forbidden) when verification fails. Implementations:
|
|
// - NoAuth — local default, everyone is anonymous (documented seam).
|
|
// - StaticToken — shared bearer token, constant-time compared (local/CI).
|
|
// - (later) OIDC — coreos/go-oidc verifier, JWKS cache, refresh-on-kid-miss.
|
|
// - (later) APIKey — hashed key behind a Secrets port, constant-time compare.
|
|
type Authenticator interface {
|
|
Authenticate(r *http.Request) (*Identity, error)
|
|
}
|
|
|
|
type identityKey struct{}
|
|
|
|
// IdentityFrom returns the authenticated identity, or false when the route was
|
|
// not behind RequireAuth.
|
|
func IdentityFrom(ctx context.Context) (*Identity, bool) {
|
|
id, ok := ctx.Value(identityKey{}).(*Identity)
|
|
return id, ok
|
|
}
|
|
|
|
// RequireAuth is route/group middleware that runs the Authenticator, rejecting
|
|
// the request (deny-by-default) when it fails and otherwise stashing the
|
|
// Identity in the context for handlers.
|
|
func RequireAuth(a Authenticator) Middleware {
|
|
return func(next HandlerFunc) HandlerFunc {
|
|
return func(c *Context) error {
|
|
id, err := a.Authenticate(c.r)
|
|
if err != nil {
|
|
return err // already an *Error (Unauthorized/Forbidden)
|
|
}
|
|
c.r = c.r.WithContext(context.WithValue(c.r.Context(), identityKey{}, id))
|
|
return next(c)
|
|
}
|
|
}
|
|
}
|
|
|
|
// NoAuth treats every caller as anonymous. The explicit local default — never
|
|
// select it for a protected environment.
|
|
type NoAuth struct{}
|
|
|
|
func (NoAuth) Authenticate(*http.Request) (*Identity, error) {
|
|
return &Identity{Subject: "anonymous"}, nil
|
|
}
|
|
|
|
// StaticToken authenticates a single shared bearer token in constant time
|
|
// (crypto/subtle) to avoid leaking the token via response timing. For local/CI
|
|
// and internal service-to-service; real principals use OIDC/API-key later.
|
|
type StaticToken struct{ token string }
|
|
|
|
// NewStaticToken builds a StaticToken; an empty token rejects every request.
|
|
func NewStaticToken(token string) StaticToken { return StaticToken{token: token} }
|
|
|
|
func (s StaticToken) Authenticate(r *http.Request) (*Identity, error) {
|
|
const prefix = "Bearer "
|
|
h := r.Header.Get("Authorization")
|
|
if s.token == "" || !strings.HasPrefix(h, prefix) {
|
|
return nil, Unauthorized("missing or malformed bearer token")
|
|
}
|
|
got := strings.TrimPrefix(h, prefix)
|
|
if subtle.ConstantTimeCompare([]byte(got), []byte(s.token)) != 1 {
|
|
return nil, Unauthorized("invalid token")
|
|
}
|
|
return &Identity{Subject: "static-token"}, nil
|
|
}
|