Files
certctl/scripts/ci-guards
shankar0123 9c679a5960 auth-bundle-2 Phase 5: OIDC + session HTTP surface (13 endpoints),
pre-login store, OpenID Connect Back-Channel Logout 1.0, cookieAuth
scheme, 7 new auth permissions, CI guard, handler tests

Phase 5 of the bundle puts the Phase 3 OIDC service + Phase 4 session
service on the wire. 13 HTTP endpoints split into three logical groups:

Public OIDC handshake (auth-exempt; protocol-mediated):
  GET  /auth/oidc/login?provider=<id>  -> 302 to IdP authorization URL
                                          + sets certctl_oidc_pending cookie
                                          (10-min TTL, Path=/auth/oidc/,
                                          SameSite=Lax)
  GET  /auth/oidc/callback?code=...&state=... -> consume pre-login row,
                                          run Phase 3's 11-step token
                                          validation, mint post-login
                                          session, 302 to dashboard
  POST /auth/oidc/back-channel-logout  -> OpenID Connect BCL 1.0 — IdP
                                          POSTs logout_token JWT; certctl
                                          validates signature against IdP
                                          JWKS via Phase 3 alg allow-list,
                                          required claims (iss/aud/iat/jti/
                                          events; exactly one of sub/sid;
                                          nonce ABSENT per spec §2.4),
                                          revokes matching sessions,
                                          returns 200 with
                                          Cache-Control: no-store
  POST /auth/logout                    -> revoke caller's session

Session management (RBAC-gated auth.session.*):
  GET    /api/v1/auth/sessions         -> auth.session.list (own / all)
  DELETE /api/v1/auth/sessions/{id}    -> auth.session.revoke (own bypass)

OIDC provider + group-mapping CRUD (RBAC-gated auth.oidc.*):
  GET    /api/v1/auth/oidc/providers              -> auth.oidc.list
  POST   /api/v1/auth/oidc/providers              -> auth.oidc.create
                                                     (client_secret encrypted
                                                     at rest via
                                                     internal/crypto.EncryptIfKeySet)
  PUT    /api/v1/auth/oidc/providers/{id}         -> auth.oidc.edit
  DELETE /api/v1/auth/oidc/providers/{id}         -> auth.oidc.delete
                                                     (refused via
                                                     ErrOIDCProviderInUse → 409
                                                     when users authenticated
                                                     via this provider)
  POST   /api/v1/auth/oidc/providers/{id}/refresh -> auth.oidc.edit
                                                     (re-runs IdP downgrade
                                                     defense via
                                                     OIDCService.RefreshKeys)
  GET    /api/v1/auth/oidc/group-mappings         -> auth.oidc.list
  POST   /api/v1/auth/oidc/group-mappings         -> auth.oidc.edit
  DELETE /api/v1/auth/oidc/group-mappings/{id}    -> auth.oidc.edit

Migration 000037 ships:

  - oidc_pre_login_sessions table (10-min absolute TTL, FK CASCADE on
    oidc_provider_id, FK RESTRICT on signing_key_id; index on
    absolute_expires_at for the GC sweep);
  - 7 new permissions seeded into r-admin only:
      auth.session.list, auth.session.list.all, auth.session.revoke,
      auth.oidc.list, auth.oidc.create, auth.oidc.edit, auth.oidc.delete

CanonicalPermissions extended in lockstep at internal/domain/auth/
validate.go.

Pre-login machinery:

  - internal/repository/oidc.go gains PreLoginRepository interface +
    PreLoginSession struct + ErrPreLoginNotFound / ErrPreLoginExpired
    sentinels.
  - internal/repository/postgres/oidc_prelogin.go ships the impl;
    LookupAndConsume uses DELETE ... RETURNING for atomic single-use.
  - internal/auth/oidc/prelogin.go is the PreLoginAdapter that bridges
    the OIDC service's Phase 3 PreLoginStore interface to the new
    repository, signing the cookie value under the active
    SessionSigningKey via the same v1.<id>.<key>.<HMAC> wire format
    Phase 4 uses for post-login cookies. Defense-in-depth: the
    pre-login `pl-` prefix is enforced by ParseCookieValue(prefix);
    a stolen pre-login cookie cannot be replayed against the
    post-login Validate path (pinned by
    TestService_Validate_RejectsPreLoginCookieAtPostLoginGate).

Session package extension:

  - internal/auth/session/service.go gains exported SignCookieValue,
    ParseCookieValue (with caller-supplied id-1 prefix), ComputeCookieHMAC,
    DecryptKeyMaterial wrappers so the OIDC pre-login adapter shares
    the same length-prefixed HMAC math without code duplication.
  - parseCookie no longer hardcodes the `ses-` prefix check (moved to
    Validate as defense-in-depth; pre-login cookie verification uses
    the `pl-` prefix via ParseCookieValue).

Cookie attributes (all Phase 5 endpoints honor CERTCTL_SESSION_SAMESITE
+ Secure=true via SessionCookieAttrs from Phase 4 config):

  - certctl_oidc_pending: Path=/auth/oidc/, MaxAge=600s, SameSite=Lax
    (cannot be Strict because the IdP-initiated callback is a top-level
    navigation from a different origin).
  - certctl_session: Path=/, Expires=8h, SameSite=Lax|Strict, HttpOnly.
  - certctl_csrf: Path=/, Expires=8h, HttpOnly=false (intentional —
    GUI must read it to echo into X-CSRF-Token header).

Audit logging on every mutating operation (event_category="auth"):

  auth.oidc_login_succeeded / failed / unmapped_groups
  auth.oidc_back_channel_logout / failed
  auth.session_revoked
  auth.oidc_provider_{created,updated,deleted,refreshed}
  auth.group_mapping_{added,removed}

OpenAPI updates:

  - cookieAuth security scheme added to api/openapi.yaml under
    components.securitySchemes (apiKey / cookie / certctl_session).
  - The 13 Phase 5 routes are added to SpecParityExceptions with a
    deferral note: full per-endpoint OpenAPI rows land in a follow-on
    commit alongside the GUI work (Phase 8) so the ergonomic shape can
    be validated against the live GUI client.

CI guard: scripts/ci-guards/N-bundle-2-security-empty-preserved.sh
asserts api/openapi.yaml has ≥ 14 'security: []' occurrences (the
pre-Bundle-2 baseline). Reducing the count below 14 would silently
force a Bearer-or-cookie requirement onto an endpoint that legitimately
runs without certctl-issued credentials; the guard fires before that
regression lands.

Handler tests (internal/api/handler/auth_session_oidc_test.go):

  - All 6 prompt-mandated negative cases:
      BCL with missing events claim -> 400
      BCL with nonce present -> 400 (per spec §2.4)
      BCL with sig signed by an unknown key -> 400
      Callback with replayed state -> 400
      Callback with PKCE verifier mismatch -> 400
      Callback with expired pre-login row -> 400
  - Plus happy paths for every endpoint, edge cases (missing-cookie,
    duplicate-name, in-use-409, wrong-tenant), and the Helper-function
    coverage (peekIssuer, classifyOIDCFailure, defaultIfBlank,
    defaultIntIfZero, clientIPFromRequest, encryptClientSecret).

Coverage on internal/api/handler/auth_session_oidc.go: 80.9% per-function
(above the Phase 5 spec's ≥ 80% floor).

Server wiring (cmd/server/main.go):

  Wired AFTER sessionService (Phase 4) so the OIDC PreLoginAdapter can
  sign pre-login cookies under the active SessionSigningKey:
    oidcProviderRepo + oidcMappingRepo + oidcUserRepo + oidcPreLoginRepo
    -> preLoginAdapter -> oidcService -> authSessionOIDCHandler.
  sessionMinterAdapter shim bridges *session.Service.Create to the
  oidcsvc.SessionMinter port the OIDC service consumes.

Router wiring (internal/api/router/router.go):

  4 public OIDC routes via direct r.mux.Handle (auth-exempt; pinned in
  AuthExemptRouterRoutes); 9 RBAC-gated routes via r.Register +
  rbacGate(checker, perm, h). Routes only register when
  reg.AuthSessionOIDC != nil so pre-Phase-5 builds skip the block
  entirely.

Verifications: gofmt clean, go vet clean across all touched packages,
go test -short -count=1 green across internal/api/handler (74 tests +
new Phase 5 batch), internal/api/router (parity + auth-exempt
allowlist), internal/auth/oidc + session (no regressions), full domain
+ scheduler + config sweeps green, ci-guard
N-bundle-2-security-empty-preserved.sh green (17 ≥ 14 baseline).
2026-05-10 06:08:27 +00:00
..
2026-05-05 18:18:29 +00:00
2026-05-05 18:18:38 +00:00

scripts/ci-guards/ — Regression-guard scripts

Each <id>.sh script in this directory pins one closed audit finding from regressing. CI runs the full set on every push via the Regression guards step in .github/workflows/ci.yml. Operators can run any script locally:

bash scripts/ci-guards/G-3-env-docs-drift.sh

Contract

Every script in this directory MUST:

  1. Be exit-code 0 on a clean repo (no regression present).
  2. Be exit-code non-zero on regression, with a ::error:: annotation prefix so PR reviewers see the failing line in the GitHub Actions UI.
  3. Be runnable from repo root via bash scripts/ci-guards/<id>.sh with NO arguments and NO env-var requirements. The CI loop step (for g in scripts/ci-guards/*.sh; do bash "$g"; done) iterates every .sh here without args; any script that requires an arg or env var WILL fail in that loop.
  4. Carry a head-comment block matching the in-source justification from the original ci.yml entry: the audit-finding reference, the closure rationale, the exempt-surface list (if any).
  5. Use set -e early to fail-fast on internal command errors.
  6. Produce no output on the happy path beyond a final echo "<id>: clean." confirmation line.

Helpers vs guards

Scripts that consume input artifacts (a test-output log, a coverage.out file) or env vars (PR_NUMBER, GH_TOKEN) are HELPERS, not guards. They live in scripts/, NOT scripts/ci-guards/.

Current helpers:

  • scripts/vendor-e2e-skip-check.sh — consumes test-output.log arg from the deploy-vendor-e2e job
  • scripts/coverage-pr-comment.sh — consumes coverage.out + PR_NUMBER + GH_TOKEN env from the go-build-and-test job
  • scripts/check-coverage-thresholds.sh — consumes coverage.out
    • .github/coverage-thresholds.yml
  • scripts/qa-doc-part-count.sh + scripts/qa-doc-seed-count.sh — invoked via make verify-docs pre-tag, not in CI

Adding a new guard

  1. Drop a new <id>.sh in this directory with the head-comment block describing the audit finding it closes.
  2. Make it executable: chmod +x scripts/ci-guards/<id>.sh.
  3. Verify it fails on a deliberate regression and passes on clean repo.
  4. CI auto-picks up new scripts via the for g in scripts/ci-guards/*.sh loop in the Regression guards step — no ci.yml change required.

The 22 guards in this directory

ID Finding Catches
G-1-jwt-auth-literal G-1 JWT silent auth downgrade "jwt" literal in additive auth-type surfaces
L-001-insecure-skip-verify L-001 unjustified InsecureSkipVerify InsecureSkipVerify: true without //nolint:gosec
H-001-bare-from H-001 (CWE-829) tag-swap attack Bare FROM line without @sha256 digest pin
M-012-no-root-user M-012 (CWE-250) container-as-root Dockerfile missing terminal USER <non-root>
H-009-readme-jwt H-009 README JWT advertising README.md re-introducing JWT-as-supported claim
G-2-api-key-hash-json G-2 cat-s5-apikey_leak api_key_hash in JSON-emitting surface
U-2-plaintext-healthcheck U-2 healthcheck protocol mismatch Plaintext http:// in HEALTHCHECK directive
U-3-migration-mount U-3 seed initdb schema drift Migration file mounted into postgres initdb
D-1-D-2-statusbadge-phantom D-1 + D-2 dead keys + TS phantoms StatusBadge dead keys + 5 Certificate / 5 Agent / 1 Issuer / 1 Notification phantom fields
L-1-bulk-action-loop L-1 client-side bulk loops for ... await triggerRenewal/updateCertificate in CertificatesPage
B-1-orphan-crud B-1 orphan-CRUD client fns 8 update/create/delete fns lose their page consumer
S-2-strings-contains-err S-2 brittle error-dispatch strings.Contains(err.Error(), "not found"|"violates foreign key") in handlers
G-3-env-docs-drift G-3 env-var docs drift CERTCTL_* env var defined OR documented but not both
test-naming-convention I-001-extended func TestXxx (lowercase first letter) — Go silently skips
S-1-hardcoded-source-counts S-1 stale numeric prose Hardcoded "N issuer connectors" / "N MCP tools" in README + docs
P-1-documented-orphan-fns P-1 documented orphans 16 read-fn names removed from client.ts exports
T-1-frontend-page-coverage T-1 untested frontend pages New page in web/src/pages/ without sibling .test.tsx and not on the deferred allowlist
bundle-8-L-015-target-blank-rel-noopener L-015 (CWE-1022) reverse-tabnabbing target="_blank" without rel="noopener noreferrer"
bundle-8-L-019-dangerously-set-inner-html L-019 (CWE-79) XSS dangerouslySetInnerHTML outside safeHtml.ts
bundle-8-M-009-bare-usemutation M-009 + M-029 mutation contract Bare useMutation() outside useTrackedMutation wrapper
H-1-encryption-key-min-length H-1 closure follow-up (post-Phase-5 surfacing) CERTCTL_CONFIG_ENCRYPTION_KEY literal in any deploy/docker-compose*.yml shorter than the 32-byte floor enforced by internal/config/config.go::Validate()
test-compose-scep-coherence post-Phase-5 surfacing of dead SCEP test config CERTCTL_SCEP_ENABLED=true in test compose without (a) a CI job that runs the SCEP integration test, (b) the ra.crt + ra.key + intune_trust_anchor.pem fixtures committed to deploy/test/fixtures/, AND (c) the matching volume mount

Guards explicitly NOT here

  • QA-doc Part-count drift + QA-doc seed-count drift — these protect docs-the-operator-reads, not anything the product depends on. Moved to make verify-docs (operator runs pre-tag, not on every push). See the ci-pipeline-cleanup spec, Phase 11.

Running the full set locally

for g in scripts/ci-guards/*.sh; do
  echo "=== $(basename "$g") ==="
  bash "$g" || echo "  FAILED"
done