Production hardening II Phase 1.
The OCSP responder previously ignored the request's nonce extension
entirely, leaving relying parties vulnerable to replay attacks. RFC
6960 §4.4.1 defines the OPTIONAL id-pkix-ocsp-nonce extension (OID
1.3.6.1.5.5.7.48.1.2): when present in the request, the responder
MUST echo the same value in the response; when absent, no nonce in
the response (back-compat with relying parties that don't send one).
NEW internal/service/ocsp_nonce.go: ParseOCSPRequestNonce walks raw
DER (golang.org/x/crypto/ocsp.Request doesn't expose the request's
extensions field — the library only exposes IssuerNameHash +
IssuerKeyHash + SerialNumber). Returns one of three states:
- (nil, false, nil) — no nonce extension in request
- (nonce, true, nil) — well-formed nonce, ≤ MaxOCSPNonceLength (32)
- (nil, false, ErrOCSPNonceMalformed) — empty or oversized
NEW internal/service/ocsp_counters.go: sync/atomic counter table for
OCSP request lifecycle (request_get/post, request_success/invalid,
nonce_echoed, nonce_malformed, rate_limited, ...). Mirrors the EST/
SCEP counter pattern; Phase 8 wires these into /metrics/prometheus.
CertSrv types extended:
- internal/connector/issuer/interface.go::OCSPSignRequest gains
Nonce []byte field.
- internal/service/renewal.go::OCSPSignRequest (the service-layer
duplicate used by ca_operations.go) gains the same field.
- internal/service/issuer_adapter.go bridges the two.
Service path: CAOperationsSvc.GetOCSPResponseWithNonce(ctx, issuerID,
serialHex, nonce) is the new entry point that plumbs the nonce
through every signing site (good / revoked / unknown / short-lived).
The legacy GetOCSPResponse becomes a nil-nonce wrapper for back-
compat — every existing caller (tests, the GET handler) sees no
behavior change.
CertificateService gains the same WithNonce variant; the handler
interface adds it to the contract. MockCertificateService in tests
extended with the new method (delegates to the legacy fn when no
override is set, so existing tests that don't care about the nonce
keep working).
Local issuer's SignOCSPResponse appends the id-pkix-ocsp-nonce
extension (non-Critical per RFC 6960 §4.4) to the response template's
ExtraExtensions when req.Nonce != nil. The extnValue is the nonce
bytes wrapped in an OCTET STRING per RFC 6960 §4.4.1.
POST OCSP handler (HandleOCSPPost):
- After ocsp.ParseRequest succeeds, calls ParseOCSPRequestNonce on
the raw body to extract the optional nonce.
- On ErrOCSPNonceMalformed (empty or > 32 bytes): writes an
'unauthorized' OCSP response (status 6 per RFC 6960 §2.3) using
the canonical ocsp.UnauthorizedErrorResponse from x/crypto/ocsp.
Does NOT echo malicious bytes back.
- On well-formed nonce: passes it through GetOCSPResponseWithNonce.
- On no nonce: nil passed through; back-compat preserved.
GET OCSP handler unchanged — the GET form has no body to carry a
nonce extension.
6 new tests in internal/service/ocsp_nonce_test.go pin every
documented failure mode + the 32-byte boundary. The test fixture
builds an OCSPRequest via golang.org/x/crypto/ocsp.CreateRequest then
splices in a [2] EXPLICIT Extensions element by hand (the library
doesn't expose extension construction either).
Pre-commit verification: gofmt clean, go vet clean across affected
packages, go test -short -count=1 green for service/ + handler/ +
connector/issuer/local/. No new env vars introduced (Phase 1 is
always-on per RFC; no operator opt-out).
Phase 4 (final phase) of the CRL/OCSP responder bundle. Closes the
backend slice; HTTP layer is now production-ready for relying parties.
What landed:
* POST /.well-known/pki/ocsp/{issuer_id} (handler.HandleOCSPPost)
- Accepts binary application/ocsp-request body per RFC 6960 §A.1.1
- Tolerant of missing Content-Type (some clients omit); validates
via ocsp.ParseRequest, returns 400 on malformed
- Returns 415 on explicit wrong Content-Type
- Reuses the existing service path (h.svc.GetOCSPResponse) — the
only new logic is body decoding + serial-from-OCSPRequest extraction
- GET form preserved unchanged for ad-hoc curl + human URL paths
- Auth-exempt under /.well-known/pki/ prefix (already in
AuthExemptDispatchPrefixes — no router changes for that)
- 7 new tests: success, method-not-allowed, wrong content-type,
missing content-type accepted, malformed body, missing issuer,
service error propagation
* router.go: r.Register("POST /.well-known/pki/ocsp/{issuer_id}", ...)
* CertificateService.GenerateDERCRL — cache-aware:
- New SetCRLCacheSvc(svc) setter (matches existing SetCAOperationsSvc
pattern — optional dep)
- When wired, GenerateDERCRL calls crlCacheSvc.Get → cheap DB read
on cache hit, singleflight-coalesced regen on miss
- When unwired, falls back to historical caSvc.GenerateDERCRL path
- GET /.well-known/pki/crl/{issuer_id} handler unchanged — calls
the same service method, gets cache benefit transparently when
the cache service is wired in cmd/server/main.go
Coverage: handler 79.8% (floor 75), service unchanged, scheduler 78%.
What's deferred (intentional scope cut for this session):
* cmd/server/main.go wiring of CRLCacheService + responder service
setters into the local issuer factory + scheduler. The wiring is
mechanical (NewCRLCacheService + scheduler.SetCRLCacheService call
in the existing wiring block); deferring keeps this commit focused
on the responder + cache primitives. Operator can wire when ready.
* Phase 5 (GUI), Phase 6 (e2e test against kind), Phase 7 (release
prep) — separate follow-up sessions.
* OCSP cache integration: today's GET/POST OCSP path goes through
the on-demand SignOCSPResponse (already cheap with the dedicated
responder cert from Phase 2). A cached-OCSP path is V3-Pro polish.
The bundle's V2 backend slice (Phases 0-4) is complete. All 4 phases
shipped 4 commits + 1 amend on this branch. CI will validate the
testcontainers repository tests on push.
Closes one 2026-04-24 audit finding (P2):
- cat-s6-efc7f6f6bd50: 30 strings.Contains(err.Error(), ...) sites
in internal/api/handler/ — brittle to repository-layer message
changes, untyped against the actual failure mode.
Approach (Option B from prompt design notes):
- New typed sentinels in internal/repository/errors.go:
ErrNotFound, ErrForeignKeyConstraint
IsForeignKeyError(err) helper (the only place substring
matching at the lib/pq boundary is allowed; isolates the
DB-driver string knowledge to one function).
- New typed sentinel in internal/domain/errors.go:
ErrValidation (reserved for future per-entity validation
wrappers; not yet used by all handlers).
- 49 sites in internal/repository/postgres/*.go updated to wrap
sql.ErrNoRows-derived errors via fmt.Errorf("...: %w",
repository.ErrNotFound).
- 18 not-found handler sites + 2 FK-constraint handler sites
refactored to errors.Is(err, repository.ErrNotFound) /
repository.IsForeignKeyError(err).
- 23 inline `fmt.Errorf("X not found")` test fixtures across
handler tests rewrapped to wrap repository.ErrNotFound.
- test_utils.go::ErrMockNotFound rewrapped to wrap
repository.ErrNotFound; renewal_policy.go closure docblock
updated to reflect the new convention.
- integration test mockJobRepository.Get wraps repository.ErrNotFound.
CI regression guardrail:
- .github/workflows/ci.yml::"Forbidden strings.Contains(err.Error())
regression guard (S-2)" greps for the three patterns ("not found",
"violates foreign key", "RESTRICT") under internal/api/handler/
and fails the build on regression.
Verification:
- go build ./... — clean
- go vet ./... — clean
- go test ./... -short -count=1 — all packages pass (handler +
repository + service + integration)
- golangci-lint v2.11.4 run ./... — 0 issues
- S-2 guardrail dry-run on post-fix tree → empty (good)
- All sibling guardrails (S-1, G-3, D-1+D-2, B-1, L-1, H-1, C-1, F-1, P-1) pass
Audit findings closed:
- cat-s6-efc7f6f6bd50 (P2)
Deferred follow-ups:
- 6 domain-specific substring patterns still inline in handlers
("cannot approve", "cannot reject", "cannot be parsed",
"no certificates found", "challenge password", "invalid"/
"required" validation chains in profiles + agent_groups). Each
needs its own typed sentinel, scoped per service. Documented
by the S-2 CI guardrail's allowlist for closure-comments only.
- Per-entity not-found sentinels (Option A — ErrCertificateNotFound,
ErrAgentNotFound, etc.) deferred. Generic ErrNotFound covers the
current dispatch needs; per-entity precision would let handlers
return entity-aware error bodies without a domain.Type field,
but not blocking.
Closes the remaining P1 gaps from coverage-gap-audit.md (M-001/M-002/M-003/M-006)
on top of the C-001/C-002 ownership + agent-FK contract fixes landed in
a53a4b8. The work lands as a single commit spanning server, docs, tests,
and the React client.
M-002 — Named API keys with per-key actor propagation
* Migration 000014 adds the 'api_keys' table (id, name, hash,
principal, role, created_at, last_used_at, disabled_at) so every
credential carries an identifiable principal instead of the
opaque 'anonymous'/'api-key' sentinel.
* Auth middleware now rotates through configured keys, performs
constant-time hash comparison, stamps 'last_used_at', and emits
an actor struct via contextWithActor(). The audit middleware,
bulk-revocation handler, approval handlers, and MCP tool layer
now read the principal off the context and persist it on every
audit_events row.
* Regression coverage:
- internal/api/middleware/audit_test.go — actor propagation,
principal redaction for disabled keys, anonymous fallback for
unauthenticated endpoints.
- internal/api/handler/bulk_revocation_handler_test.go,
job_handler_test.go — principal-on-audit assertions.
M-003 — Authorization gates (Phase B)
* Approval handler rejects self-approval / self-rejection with 403
when the actor principal equals the job's requested_by field.
* Bulk revocation is gated behind the 'admin' role; operators and
viewers receive 403.
* Regression coverage:
- internal/service/job_test.go — TestApproveJob_NotSelf,
TestRejectJob_NotSelf.
- internal/api/handler/bulk_revocation_handler_test.go —
TestBulkRevoke_RequiresAdmin, TestBulkRevoke_AdminSucceeds.
M-006 — RFC-compliant CRL/OCSP on the unauthenticated .well-known mux
* Per RFC 8615, relying parties cannot reasonably be asked to
authenticate against the issuing certctl instance to retrieve
revocation material. CRL and OCSP move off the authenticated
'/api/v1/crl*' and '/api/v1/ocsp/*' paths onto:
GET /.well-known/pki/crl/{issuer_id}
Content-Type: application/pkix-crl (RFC 5280 §5)
GET /.well-known/pki/ocsp/{issuer_id}/{serial}
Content-Type: application/ocsp-response (RFC 6960)
* Non-standard JSON CRL shape is removed; only DER is served.
* Short-lived certificate exemption (profile TTL < 1h → skip
CRL/OCSP) is preserved; the response simply omits the serial.
* Routes are registered on the unauthenticated 'finalHandler' mux
in cmd/server/main.go alongside EST ('/.well-known/est/*') and
SCEP ('/scep'). Legacy authenticated paths return 404.
* Regression coverage:
- internal/api/handler/certificate_handler_test.go — content
type, DER parseability, 404 for unknown issuer.
- internal/api/handler/adversarial_path_test.go — unauthenticated
access asserted for CRL, OCSP, EST, SCEP.
- internal/api/router/router_test.go — route-table assertion
that '.well-known/pki/*', '.well-known/est/*', and '/scep' are
mounted on the unauthenticated branch.
M-001 — Auto-closed by M-002
EST and SCEP were already registered on the unauthenticated
'finalHandler' mux; the router comment at
internal/api/router/router.go:247 now matches reality. The
adversarial-path tests above lock the behavior in.
Verification (all gates green):
* go vet ./... — clean
* go build ./... — ok
* go test -short ./... (55+ packages) — all pass
* web/ : npm test (225 Vitest tests) — all pass
* web/ : npx tsc --noEmit — clean
* grep sweep for '/api/v1/(crl|ocsp)' — 13 surviving hits,
all intentional M-006 tombstone/relocation comments.
Documentation:
* coverage-gap-audit.md — status flips M-001/M-002/M-003/M-006 →
Fixed, with per-finding resolution paragraphs citing regression
test IDs. (Audit file lives outside this repo; see cowork root.)
* CLAUDE.md Project Status line updated with the auth-unification
closure note.
* docs/features.md, docs/architecture.md, docs/quickstart.md,
docs/concepts.md, docs/connectors.md, docs/test-env.md,
docs/testing-guide.md, docs/compliance-*.md, docs/demo-advanced.md
— refreshed for the new '.well-known/pki/*' namespace and named
API keys.
* api/openapi.yaml — documents the new unauthenticated endpoints
and removes the legacy '/api/v1/crl*' + '/api/v1/ocsp/*' paths.
.gitignore: adds '/.gocache/' and '/.gomodcache/' for the session-
scoped Go caches so they never enter the tree.
C-001 — CreateCertificate was server-accepted with null owner_id,
team_id, renewal_policy_id because the GUI neither collected the fields
nor enforced them, even though the backend's ManagedCertificate schema
and handler contract treat them as required. Fix the contract at all
four layers:
- web/src/pages/CertificatesPage.tsx: replace owner_id/team_id free-
text inputs with <select> elements fed by getOwners/getTeams/
getPolicies queries; mark all three required; gate the Create
button on owner_id + team_id + renewal_policy_id being set.
- internal/api/handler/certificates.go: ValidateRequired for
owner_id, team_id, renewal_policy_id on CreateCertificate so the
handler returns HTTP 400 with the offending field name before the
service layer is reached.
- internal/mcp/types.go: drop ',omitempty' from
CreateCertificateInput.RenewalPolicyID so the MCP schema reflects
the required contract; Update inputs keep partial-update semantics.
- api/openapi.yaml: 'required: [name, common_name, renewal_policy_id,
issuer_id, owner_id, team_id]' was already present on the Create
schema; clarified DeploymentTarget.agent_id description to note the
FK contract.
C-002 — CreateTargetWizard accepted an empty or bogus agent_id and the
service inserted directly, producing a Postgres 23503 FK-violation that
bubbled out as a generic HTTP 500. The FK itself (migration 000001 line
104: agent_id TEXT NOT NULL REFERENCES agents(id)) is correct; we keep
the schema strict and add validation at three layers:
- internal/service/target.go: introduce
ErrAgentNotFound sentinel and pre-validate agent_id in
TargetService.CreateTarget — empty string returns
'agent_id is required'; a nonexistent id returns the full
'referenced agent does not exist: <id>' error. Both wrap
ErrAgentNotFound via fmt.Errorf %w so callers can use errors.Is.
- internal/api/handler/targets.go: ValidateRequired on agent_id; map
errors.Is(err, service.ErrAgentNotFound) to HTTP 400 instead of
letting it fall through to the generic 500 branch.
- internal/mcp/types.go: drop ',omitempty' from
CreateTargetInput.AgentID to match the required contract.
- web/src/pages/TargetsPage.tsx: replace the free-text Agent ID input
with a <select> populated from getAgents(); include agent in the
canProceedToReview gate so Next is disabled until an agent is
chosen.
Regression coverage (21 new subtests total):
- TestCreateCertificate_MissingRequiredField_Returns400 — 6 subtests,
one per required field, each proves the handler guard fires before
the mock service is called.
- TestCreateTarget_MissingAgentID_Returns400 — handler guard.
- TestCreateTarget_NonexistentAgent_Returns400 — pins the
ErrAgentNotFound -> 400 translation.
- TestTargetService_CreateTarget_MissingAgentID — errors.Is sentinel.
- TestTargetService_CreateTarget_NonexistentAgentID — errors.Is.
- The existing TestTargetService_CreateTarget_Success, along with
TestCreateTarget_{MissingName,MissingType,NameTooLong}_* handler
tests, were updated to seed a real agent or include agent_id in
the request body so the happy paths still run cleanly.
Gates (Phase 4):
- go build/vet/test/race: green
- go test -cover: internal/service 68.7% (gate 55%),
internal/api/handler 78.9% (gate 60%)
- golangci-lint on service+handler+mcp: 0 issues
- govulncheck: no reachable vulns
- tsc --noEmit: clean
- vitest: 223/223 passing
See cowork/certctl-coverage-gap-audit.md entries C-001 and C-002.
Collapses CertificateService, RevocationSvc, and CAOperationsSvc to
ctx-accepting method signatures. Removes context.Background() synthesis
at 24 internal call sites across certificate.go, revocation_svc.go, and
ca_operations.go.
- Primary repo calls inherit request cancellation via the passed ctx.
- Audit and notification dispatches use context.WithoutCancel(ctx) so
they survive client disconnect.
- Collapses TriggerRenewal/TriggerRenewalWithActor,
TriggerDeployment/TriggerDeploymentWithActor, and
RevokeCertificate/RevokeCertificateWithActor sibling pairs into single
canonical ctx-accepting methods (decisions D-1, D-2).
Handlers pass r.Context(). Mocks and tests updated to match new
signatures. No HTTP surface change, no OpenAPI change.
PR 1 of 6 in the M-2 remediation chain. Master green at this commit.
Refs: certctl-audit-report.md M-2 (L143, L224)
Root cause: certificate_versions.csr_pem is nullable in the schema but
Go code scanned it into a plain string. Used sql.NullString in
ListVersions and GetLatestVersion to handle NULL values correctly.
Also includes: partial update fetch-merge-update pattern to prevent FK
violations, nil directory guard in discovery service, diagnostic slog
logging in handlers, export handler 422 for unparseable PEM, OpenAPI
spec corrections, MCP tool description improvements, and test fixes.
Rewrites the Release Sign-Off section in testing-guide.md to individual
test-level granularity (320 rows) with smoke test results audited and
checked off (121 pass, 5 skip, 194 manual remaining).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Service tests: newRevocationTestService() was missing SetIssuerRegistry(),
causing all 8 CRL/OCSP tests to fail with "issuer registry not configured".
Handler tests: CRL tests used /api/v1/issuers/{id}/crl but handler parses
/api/v1/crl/{id}. OCSP tests used query string ?serial=X but handler
expects path param /api/v1/ocsp/{id}/{serial}. Fixed all 9 test URLs.
All issues pre-date CI on v2-dev — introduced during M15b.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Backend:
- Embedded OCSP responder: GET /api/v1/ocsp/{issuer_id}/{serial} returns
signed OCSP responses (good/revoked/unknown) using CA key
- DER-encoded X.509 CRL: GET /api/v1/crl/{issuer_id} returns proper DER CRL
signed by issuing CA with 24h validity window
- Short-lived cert exemption: certs with profile TTL < 1 hour skip CRL/OCSP
(expiry is sufficient revocation for ephemeral workloads)
- Extended issuer connector interface with GenerateCRL and SignOCSPResponse
- Local CA implements full CRL/OCSP signing; ACME and step-ca return
appropriate "use native endpoint" errors
- IssuerConnectorAdapter bridges new methods between layers
Frontend:
- Revoke button on certificate detail page with RFC 5280 reason modal
- Revocation banner with reason display and timestamp
- Revocation status indicators in lifecycle section
- "Revoked" filter option in certificates list
- API client: revokeCertificate() function and Certificate type extensions
Tests: ~31 new tests across connector, service, handler, and adapter layers
Docs: milestones renumbered (M13-M14, M16-M18), M15b marked complete
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Implements core revocation infrastructure: POST /api/v1/certificates/{id}/revoke
with all 8 RFC 5280 reason codes, JSON-formatted CRL at GET /api/v1/crl, webhook
and email revocation notifications, best-effort issuer notification, and immutable
revocation audit trail. Includes 48 new tests across service, handler, integration,
and domain layers (600+ total). Fixes 3 pre-existing test bugs (team_test error
matching, agent_group delete status code, team handler per_page validation).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Backend hardening:
- Fix 6 nginx.go non-constant format string build errors
- Add validation.go with hostname, PEM, and enum validators
- Apply input validation to all POST/PUT handlers (certificates,
agents, CSR, policies, teams, owners, targets, issuers)
- Fix unchecked JSON decode in TriggerDeployment handler
Frontend (Vite + React + TypeScript):
- Migrate from single-file SPA to proper build pipeline
- 7 pages: Dashboard, Certificates (list+detail), Agents, Jobs,
Notifications, Policies, Audit Trail
- TanStack Query for server state with auto-refetch intervals
- Certificate detail with version history and renewal trigger
- Job cancellation, status/type filtering, expiry countdowns
- Reusable components: DataTable, StatusBadge, ErrorState, PageHeader
- Dark theme with Tailwind CSS, sidebar nav via React Router
Server integration:
- Go server serves web/dist/ (Vite output) with SPA fallback
- Falls back to web/index.html for legacy mode
- .gitignore updated for web/node_modules/ and web/dist/
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Service layer (63 tests): certificate, agent, audit, job, notification,
policy, and renewal services with mock repositories covering threshold
alerting, deduplication, status transitions, and job processing.
Handler layer (46 tests): certificate and agent HTTP handlers using
httptest with mock service interfaces, covering success/error paths,
pagination, JSON marshaling, and path parameter extraction.
Integration (11 subtests): end-to-end certificate lifecycle test
exercising real services and Local CA issuer through HTTP API —
create cert, trigger renewal, process jobs, register agent, heartbeat,
verify audit trail.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
