CANP-WP-0006 T06: container image and smoke checks

Two-stage python:3.12-slim build with no toolchain in the runtime layer,
running non-root (uid 10001) and writing nothing to disk — its state is the
database. reference/ is a real build input, because the service delegates
validation to it so the two cannot disagree about what a valid package is.

Migrations deliberately do not run at start-up. A schema change is a deployment
step with its own rollback, not something that races between replicas.

tools/smoke.py asserts what can be known about a running service: liveness,
readiness, fleet health shape, migration revision, and that the index and
registry are queryable. stdlib only, so it runs inside the runtime image;
non-zero exit, so a deployment gate can call it directly.

Verified by running it, not by inspection: the container starts, all six checks
pass against it, the reference CLI installs a package from it over HTTP, and
the checks fail correctly against a wrong --expect-migration — so
migration-at-head is a real check rather than a decorative one. Without
--expect-migration the check can only confirm the schema is stamped at all, and
says so rather than implying it verified the head.

Cluster-level checks a rapp contract also names — NetworkPolicies present,
external secrets ready, private-Service-only, live image digest match — are
properties of the deployment and belong to rapp-canned-prompts.

The digest rapp.yaml would pin does not exist yet. The image was built locally
and verified, but never pushed; that digest exists only once the image is
published to the fleet registry, which needs credentials and is outward-facing
enough not to do unasked. The follow-on sequence for rapp-canned-prompts is
recorded in the workplan.

CANP-WP-0006 is finished. Service tests 33, reference 105.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Bjefh8NUiEiahN4JLwoSKM

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 388925@bnt-lap001
Assistant-Session: 3507023f-e0fd-4a1e-9d90-a0d4217d1502

service/.dockerignore

@@ -0,0 +1,4 @@
+**/.venv/
+**/__pycache__/
+**/*.pyc
+**/*.db

service/Dockerfile

@@ -0,0 +1,40 @@
+# syntax=docker/dockerfile:1
+
+# Build stage: resolve dependencies into a wheel-installed prefix so the runtime
+# image carries no build toolchain.
+FROM python:3.12-slim AS build
+
+ENV PIP_DISABLE_PIP_VERSION_CHECK=1 PIP_NO_CACHE_DIR=1
+WORKDIR /build
+
+# The reference implementation is a real dependency of the service: validation is
+# delegated to it so the service and the CLI cannot disagree about what a valid
+# package is.
+COPY reference/ ./reference/
+COPY service/pyproject.toml ./service/
+COPY service/src/ ./service/src/
+
+RUN python -m venv /opt/venv \
+ && /opt/venv/bin/pip install --no-cache-dir ./reference ./service
+
+FROM python:3.12-slim AS runtime
+
+# Non-root by default. The service writes nothing to disk; its state is the
+# database.
+RUN useradd --system --create-home --uid 10001 canned
+ENV PATH="/opt/venv/bin:$PATH" \
+    PYTHONDONTWRITEBYTECODE=1 \
+    PYTHONUNBUFFERED=1
+
+COPY --from=build /opt/venv /opt/venv
+COPY service/alembic.ini /app/alembic.ini
+COPY service/migrations/ /app/migrations/
+
+WORKDIR /app
+USER canned
+EXPOSE 8000
+
+# No CMD-embedded migration: schema changes are a deployment step with its own
+# rollback, not something that races between replicas at start-up.
+CMD ["uvicorn", "canned_prompts_service.api:create_app", "--factory", \
+     "--host", "0.0.0.0", "--port", "8000"]

service/Makefile

@@ -0,0 +1,28 @@
+# Service build and verification. Run from `service/`.
+IMAGE ?= canned-prompts-service
+TAG   ?= dev
+BASE  ?= http://127.0.0.1:8000
+HEAD  := $(shell .venv/bin/alembic heads 2>/dev/null | awk '{print $$1}')
+
+.PHONY: test image smoke migrate run head
+
+test:
+	.venv/bin/python -m pytest -q
+
+image:
+	cd .. && docker build -f service/Dockerfile -t $(IMAGE):$(TAG) .
+
+head:
+	@echo $(HEAD)
+
+migrate:
+	.venv/bin/alembic upgrade head
+
+run:
+	.venv/bin/uvicorn canned_prompts_service.api:create_app --factory
+
+# The service-level half of a rapp smoke contract. Cluster-level checks
+# (NetworkPolicies, external secrets, private-Service-only) belong to
+# rapp-canned-prompts, which can call this for the rest.
+smoke:
+	python3 tools/smoke.py --base $(BASE) --expect-migration $(HEAD)

service/README.md

@@ -119,12 +119,42 @@ it cannot be attributed among several publishers. Until per-publisher identity
 exists, a claim's `owner` is documentation rather than an access decision, and
 the code says so where it matters.
 
+## Image and smoke
+
+```bash
+make image                  # builds from the repo root; reference/ is a real dependency
+make smoke BASE=http://...  # service-level checks, exits non-zero on failure
+```
+
+The image is a two-stage `python:3.12-slim` build carrying no build toolchain,
+running as a non-root user, writing nothing to disk — its state is the database.
+It deliberately does **not** run migrations on start-up: a schema change is a
+deployment step with its own rollback, not something that races between replicas.
+
+`tools/smoke.py` asserts what can be known about a *running service*: liveness,
+readiness, fleet health shape, the migration revision, and that the index and
+registry are queryable. It is stdlib-only so it runs inside the runtime image,
+and it exits non-zero so a deployment gate can call it directly.
+
+Cluster-level checks a `rapp` contract also names — NetworkPolicies present,
+external secrets ready, private-Service-only, live image digest match — are
+properties of the deployment rather than of this process, and belong to
+`rapp-canned-prompts`.
+
+`--expect-migration` matters: without it the check can only confirm the schema
+is stamped at all, and it says so rather than implying it verified the head.
+
 ## Status
 
-`CANP-WP-0006` T01–T04 are done: skeleton, health surface, tenant-keyed schema
-(migrations `0001`–`0002`), and the read and publish APIs. The HTTP registry
-client in the CLI and the container image (T05–T06) are not built yet. T06
-produces the image digest that `rapp-canned-prompts` needs to pin.
+`CANP-WP-0006` is complete: skeleton and health surface, tenant-keyed schema
+(migrations `0001`–`0002`), read and publish APIs, HTTP registries in the
+reference CLI, and a verified container image with smoke checks.
 
-Per-publisher identity is deferred, and is the main thing standing between this
-and a service that several people can publish to.
+**Not done, and needed before `rapp-canned-prompts`:** the image has been built
+and verified locally but never pushed. `rapp.yaml` pins
+`upstream_components.version` to a digest from the fleet's registry, and that
+digest only exists once the image is published — an operator action needing
+registry credentials.
+
+**Per-publisher identity** remains deferred, and is the main thing between this
+and a registry several people can publish to.

service/tools/smoke.py

@@ -0,0 +1,111 @@
+#!/usr/bin/env python3
+"""Service-level smoke checks.
+
+The checks a `rapp` smoke contract can assert about a *running service*.
+Deployment-level checks — NetworkPolicies present, external secrets ready,
+private-Service-only — are properties of the cluster, not of this process, and
+belong to `rapp-canned-prompts` rather than here.
+
+Exits non-zero if any required check fails, so a deployment gate can call it
+directly. stdlib only: this runs inside the runtime image, which carries no
+test dependencies.
+"""
+
+from __future__ import annotations
+
+import argparse
+import json
+import sys
+import urllib.error
+import urllib.request
+from dataclasses import dataclass
+
+
+@dataclass
+class Check:
+    name: str
+    ok: bool
+    detail: str
+
+
+def get(base: str, path: str, timeout: float) -> tuple[int, dict]:
+    request = urllib.request.Request(base.rstrip("/") + path, method="GET")
+    request.add_header("Accept", "application/json")
+    try:
+        with urllib.request.urlopen(request, timeout=timeout) as response:
+            return response.status, json.loads(response.read().decode("utf-8"))
+    except urllib.error.HTTPError as exc:
+        try:
+            return exc.code, json.loads(exc.read().decode("utf-8"))
+        except Exception:  # noqa: BLE001
+            return exc.code, {}
+    except Exception as exc:  # noqa: BLE001
+        return 0, {"error": str(exc)}
+
+
+def run(base: str, expect_migration: str | None, timeout: float) -> list[Check]:
+    checks: list[Check] = []
+
+    status, body = get(base, "/healthz", timeout)
+    checks.append(
+        Check("liveness-ok", status == 200 and body.get("status") == "ok",
+              f"{status} {body}")
+    )
+
+    status, body = get(base, "/readyz", timeout)
+    ready = status == 200 and body.get("ready") is True
+    checks.append(Check("readiness-ok", ready, f"{status} {body.get('detail')}"))
+
+    status, body = get(base, "/state/health", timeout)
+    checks.append(
+        Check("state-health-ok", status == 200 and body.get("status") == "ok",
+              f"{status} {body.get('db')}")
+    )
+
+    migration = body.get("migration")
+    if expect_migration:
+        checks.append(
+            Check("migration-at-head", migration == expect_migration,
+                  f"running {migration}, expected {expect_migration}")
+        )
+    else:
+        # Without an expected revision this can only confirm the schema is
+        # stamped at all — say so rather than implying it verified the head.
+        checks.append(
+            Check("migration-stamped", bool(migration),
+                  f"running {migration} (no --expect-migration given)")
+        )
+
+    status, body = get(base, "/index", timeout)
+    checks.append(Check("index-queryable", status == 200 and "entries" in body,
+                        f"{status}"))
+
+    status, body = get(base, "/packages", timeout)
+    checks.append(Check("registry-queryable", status == 200 and "packages" in body,
+                        f"{status}"))
+    return checks
+
+
+def main(argv: list[str] | None = None) -> int:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--base", default="http://127.0.0.1:8000")
+    parser.add_argument("--expect-migration", default=None,
+                        help="revision the deployed schema should be at")
+    parser.add_argument("--timeout", type=float, default=5.0)
+    args = parser.parse_args(argv)
+
+    checks = run(args.base, args.expect_migration, args.timeout)
+    width = max(len(c.name) for c in checks)
+    for check in checks:
+        print(f"{'PASS' if check.ok else 'FAIL'}  {check.name:<{width}}  {check.detail}")
+
+    failed = [c.name for c in checks if not c.ok]
+    if failed:
+        print(f"\n{len(failed)} check(s) failed: {', '.join(failed)}", file=sys.stderr)
+        return 1
+    print(f"\nall {len(checks)} checks passed")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

workplans/CANP-WP-0006-hosted-registry-service.md

@@ -4,7 +4,7 @@ type: workplan
 title: "Hosted registry and index service"
 domain: agents
 repo: canned-prompts
-status: active
+status: finished
 owner: codex
 topic_slug: practice
 created: "2026-09-06"
@@ -245,7 +245,7 @@ its only dependency. Registry responses are treated as untrusted input (§ 19):
 
 ```task
 id: CANP-WP-0006-T06
-status: todo
+status: done
 priority: medium
 state_hub_task_id: "3c008ebe-bf07-5cee-8678-7b1ed27283ef"
 ```
@@ -253,5 +253,42 @@ state_hub_task_id: "3c008ebe-bf07-5cee-8678-7b1ed27283ef"
 Dockerfile and a published image, plus the checks a `rapp` smoke contract will
 assert: health ok, migration at head, private service only, image digest match.
 
-Completing this produces the digest that `rapp-canned-prompts` needs to pin, at
-which point that repo can be created against something real.
+**Done, with one honest limit.**
+
+The image builds and was verified by running it: a two-stage `python:3.12-slim`
+build with no toolchain in the runtime layer, non-root (uid 10001), writing
+nothing to disk. All six smoke checks passed against the running container, and
+the reference CLI installed a package from it over HTTP.
+
+Migrations deliberately do **not** run at start-up. A schema change is a
+deployment step with its own rollback, not something that races between
+replicas.
+
+`tools/smoke.py` covers what can be known about a *running service* — liveness,
+readiness, fleet health shape, migration revision, index and registry
+queryable. stdlib only, so it runs inside the runtime image; non-zero exit, so
+a deployment gate can call it. Verified in both directions: it passes against
+the container and fails against a wrong `--expect-migration`, so
+`migration-at-head` is a real check rather than a decorative one.
+
+Cluster-level checks a rapp contract also names — NetworkPolicies present,
+external secrets ready, private-Service-only, live image digest match — are
+properties of the deployment and belong to `rapp-canned-prompts`.
+
+**The digest does not exist yet.** The image was built locally
+(`sha256:4878b208…`, 76 MB) but never pushed. `rapp.yaml` pins
+`upstream_components.version` to a digest from the fleet's registry, which
+exists only once the image is published — an operator action needing registry
+credentials, and outward-facing enough that it is not mine to take unasked.
+
+## Follow-on: rapp-canned-prompts
+
+Not a task in this workplan; recorded so the sequence is not lost.
+
+1. Publish the image to the fleet registry and capture its digest.
+2. Create `rapp-canned-prompts` with `ownership_repo: canned-prompts`,
+   `readiness_state: draft`, and that digest in `upstream_components`.
+3. Add the deployment-level smoke checks, calling `tools/smoke.py` for the
+   service-level half.
+4. Decide the PostgreSQL binding with `rapp-postgres` and the credential
+   broker, per the shape `rapp-sbom-nexus` uses.