Files
certctl/internal/auth/oidc/test_discovery.go
T
shankar0123 fefeccfa59 harden(oidc): relax alg-downgrade IdP-bind check to intersection-empty (Keycloak compat)
Phase-10 live-IdP smoke (Keycloak 26.x via testcontainers-go) revealed
the IdP-bind alg-downgrade check was too strict for real-world IdPs.
6 of the integration tests in internal/auth/oidc/integration_keycloak*_test.go
were failing with:

  oidc: IdP advertises weak signing algorithms (HS*/none);
  refusing to use as defense against downgrade attacks: HS256

Keycloak 26.x (and several other real-world IdPs — Auth0 when HS-mode is
enabled, some Authentik configs) advertise EVERY alg they're capable of
in the discovery doc's id_token_signing_alg_values_supported field, even
when the realm only signs with RS256 in practice. Pre-fix the IdP-bind
check refused on ANY HS* or 'none' advertisement → no real Keycloak deploy
could ever bind a provider row, hence the integration-test failures.

The strict-deny check was defense-in-depth on top of the load-bearing
per-token alg-pin at sig-verify time (isDisallowedAlg, service.go L1177):
that check rejects every ID token whose JWS header carries an alg outside
DefaultAllowedAlgs, regardless of what the discovery doc advertises.
A forged HS256 token signed with the IdP's RS256 pubkey as HMAC secret
is rejected at sig-verify time → the actual algorithm-confusion attack
is closed by the per-token pin, NOT by the discovery-doc check.

Fix: relax the IdP-bind check to refuse only when the intersection of
advertised vs DefaultAllowedAlgs is EMPTY (the pathological all-weak-alg
IdP case). Keycloak (RS256 + HS256 advertised) now binds successfully;
an HS-only IdP still fails closed.

Changes:
- internal/auth/oidc/service.go: rewrite the alg-check loop at L1067 in
  getOrLoad / RefreshKeys to compute the intersection set; refuse only
  when no acceptable alg is advertised. ErrIdPDowngradeAdvertised
  docstring updated to reflect new contract. DefaultAllowedAlgs
  docstring + the package-level design-comment block at L40-72 updated
  with v2.1.0-relaxed semantics callouts.
- internal/auth/oidc/test_discovery.go: TestDiscovery dry-run validator
  rewritten to surface HS*/none alongside RS* as an informational note
  ('note: IdP advertises weak algorithms %v alongside acceptable ones')
  rather than a hard-fail error. HS-only / none-only still hard-fails.
- internal/auth/oidc/service_test.go: TestService_IdPDowngradeDefense_*
  tests updated. Renamed:
  - RejectsHSAdvertised → RS256PlusHS256_BindsSuccessfully (positive)
  - RejectsNoneAdvertised → RejectsHSOnlyAdvertised (intersection-empty)
  - RefreshKeys_CatchesPostLoadDowngrade rotated to HS-only post-load
- internal/auth/oidc/coverage_fill_test.go: TestTestDiscovery_AlgDowngradeDetected
  split into _HS256AlongsideRS256_BindsWithNote (positive, asserts note
  but no hard-fail) + _HSOnly_StillTrips_HardFail (intersection-empty).
- docs/operator/auth-threat-model.md: OIDC token-validation alg-allow-list
  section rewritten to call out the load-bearing-defense hierarchy
  (per-token pin first, IdP-bind check defense-in-depth) and document
  the v2.1.0 relaxation rationale.
- CHANGELOG.md: ### Security entry under Unreleased.

Verify: go test ./internal/auth/oidc/ -short PASS; gofmt clean; go vet
clean. The Keycloak integration tests should now pass when the operator
re-runs 'make keycloak-integration-test'.
2026-05-11 15:34:59 +00:00

149 lines
6.4 KiB
Go

package oidc
// Audit 2026-05-10 MED-5 closure — dry-run validator for OIDC provider
// configuration. Lets operators verify discovery + JWKS reachability +
// alg-downgrade defense BEFORE persisting a provider row. Mirrors the
// non-persistence-touching subset of getOrLoad.
import (
"context"
"fmt"
"net/http"
gooidc "github.com/coreos/go-oidc/v3/oidc"
)
// TestDiscoveryResult is the report TestDiscovery returns. The HTTP
// layer marshals this verbatim. Each field is independently observable
// so the GUI can render a per-check status row.
//
// `Errors` collects every leg that failed; a partial-success case
// (e.g. discovery OK but alg-downgrade tripped) returns
// DiscoverySucceeded=true + a non-empty Errors slice.
type TestDiscoveryResult struct {
DiscoverySucceeded bool `json:"discovery_succeeded"`
JWKSReachable bool `json:"jwks_reachable"`
SupportedAlgValues []string `json:"supported_alg_values"`
IssParamSupported bool `json:"iss_param_supported"`
IssuerEcho string `json:"issuer_echo,omitempty"` // the iss value the IdP advertised
AuthorizationURL string `json:"authorization_url,omitempty"`
TokenURL string `json:"token_url,omitempty"`
JWKSURI string `json:"jwks_uri,omitempty"`
UserInfoEndpoint string `json:"userinfo_endpoint,omitempty"`
Errors []string `json:"errors,omitempty"`
}
// TestDiscovery runs the read-only subset of getOrLoad against a
// candidate issuer URL: fetches the discovery doc, runs the
// alg-downgrade defense, parses the RFC 9207 iss-parameter advert,
// then fetches the JWKS once to confirm reachability.
//
// The function NEVER persists anything; the caller is the
// /api/v1/auth/oidc/test endpoint that the GUI uses for dry-runs.
//
// Service-layer entry point so the handler stays HTTP-shaped only.
func (s *Service) TestDiscovery(ctx context.Context, issuerURL string) (*TestDiscoveryResult, error) {
res := &TestDiscoveryResult{}
// Step 1 — discovery. gooidc.NewProvider fetches
// `<issuer>/.well-known/openid-configuration` and runs the iss
// match check internally; on failure it returns a fmt-style
// wrapped error.
provider, err := gooidc.NewProvider(ctx, issuerURL)
if err != nil {
res.Errors = append(res.Errors, fmt.Sprintf("discovery fetch failed: %v", err))
return res, nil // Non-fatal at this layer; the response carries the per-leg failure.
}
res.DiscoverySucceeded = true
res.IssuerEcho = issuerURL
endpoint := provider.Endpoint()
res.AuthorizationURL = endpoint.AuthURL
res.TokenURL = endpoint.TokenURL
// Step 2 — parse the claims we care about from the discovery doc.
var advertised struct {
IDTokenSigningAlgValuesSupported []string `json:"id_token_signing_alg_values_supported"`
AuthorizationResponseIssParamSupported bool `json:"authorization_response_iss_parameter_supported"`
JWKSURI string `json:"jwks_uri"`
UserInfoEndpoint string `json:"userinfo_endpoint"`
}
if cerr := provider.Claims(&advertised); cerr != nil {
res.Errors = append(res.Errors, fmt.Sprintf("discovery claims: %v", cerr))
return res, nil
}
res.SupportedAlgValues = advertised.IDTokenSigningAlgValuesSupported
res.IssParamSupported = advertised.AuthorizationResponseIssParamSupported
res.JWKSURI = advertised.JWKSURI
res.UserInfoEndpoint = advertised.UserInfoEndpoint
// Step 3 — alg-downgrade defense (v2.1.0-relaxed semantics).
// Pre-v2.1.0 this loop appended an error for ANY HS*/none in the
// IdP's advertised list. That was strict-deny but incompatible with
// real IdPs like Keycloak 26.x which list every alg they're capable
// of, even though the realm only signs with RS256.
// New semantics: only flag the IdP if the intersection of advertised
// vs DefaultAllowedAlgs is empty (a pathological all-weak IdP). Each
// HS*/none advertisement is still surfaced as an informational note
// so operators can ask their IdP team to tighten the list, but it's
// no longer a hard fail. The per-token alg check at sig-verify time
// (isDisallowedAlg in service.go ~L1177) is the load-bearing defense.
allowedSet := make(map[string]struct{}, len(DefaultAllowedAlgs))
for _, a := range DefaultAllowedAlgs {
allowedSet[a] = struct{}{}
}
hasAcceptable := false
weak := []string{}
for _, a := range advertised.IDTokenSigningAlgValuesSupported {
if _, ok := allowedSet[a]; ok {
hasAcceptable = true
}
if _, deny := disallowedAlgs[a]; deny {
weak = append(weak, a)
}
}
if len(advertised.IDTokenSigningAlgValuesSupported) > 0 && !hasAcceptable {
res.Errors = append(res.Errors, fmt.Sprintf("alg-downgrade defense tripped: IdP advertises only weak algorithms (%v) — no acceptable alg from %v present", advertised.IDTokenSigningAlgValuesSupported, DefaultAllowedAlgs))
} else if len(weak) > 0 {
// Informational only — RS/ES present alongside HS, so the
// IdP binds successfully but the operator should know.
res.Errors = append(res.Errors, fmt.Sprintf("note: IdP advertises weak algorithms %v alongside acceptable ones — verifier-side alg pin prevents downgrade, but tightening the IdP's advertised list is recommended", weak))
}
// Step 4 — JWKS reachability. The go-oidc Verifier defers JWKS
// fetch until first token-verify; for the dry-run we explicitly
// HEAD/GET the JWKS endpoint to confirm network reachability.
if advertised.JWKSURI == "" {
res.Errors = append(res.Errors, "discovery doc omits jwks_uri")
} else if ok, herr := jwksReachable(ctx, advertised.JWKSURI); !ok {
if herr != nil {
res.Errors = append(res.Errors, fmt.Sprintf("JWKS fetch failed: %v", herr))
} else {
res.Errors = append(res.Errors, "JWKS endpoint returned non-200")
}
} else {
res.JWKSReachable = true
}
return res, nil
}
// jwksReachable issues a GET against the JWKS URI and returns ok=true
// when the response status is 2xx. Used by TestDiscovery for the
// reachability leg of the dry-run.
//
// Kept distinct from go-oidc's internal JWKS fetcher because we want
// to surface the HTTP status to the operator without requiring a
// token-verify round-trip.
var jwksReachable = func(ctx context.Context, jwksURI string) (bool, error) {
req, err := http.NewRequestWithContext(ctx, http.MethodGet, jwksURI, nil)
if err != nil {
return false, err
}
resp, err := http.DefaultClient.Do(req)
if err != nil {
return false, err
}
defer resp.Body.Close()
return resp.StatusCode >= 200 && resp.StatusCode < 300, nil
}