47eb56bd10
The previous fix (`host.docker.internal:<port>` for daemon-side push) still failed: Get "https://host.docker.internal:53958/v2/": http: server gave HTTP response to HTTPS client `host.docker.internal` is reachable from Docker Desktop's daemon VM but isn't in the daemon's default insecure-registries CIDRs (only `::1/128` and `127.0.0.0/8` are), so docker push tries HTTPS, hits a plain-HTTP registry, and refuses. The daemon.json fix (`"insecure-registries": ["host.docker.internal"]`) works but is a one-time manual step in Docker Desktop's UI — not something we can do for the user. Sidestep the daemon push entirely: 1. docker build (as before) — local layer cache makes no-change rebuilds cheap. 2. docker save the image to a per-digest tarball alongside the cached `.smolmachine`. 3. Start an ephemeral registry container on a per-session docker network, with `-p :5000` so the host can also reach it for the pack step. 4. docker run a one-shot crane container on the SAME network, mount the tarball, `crane push --insecure /img.tar <registry-container>:5000/...`. Container DNS resolves the registry on the network; `--insecure` forces plain HTTP. 5. `smolvm pack create --image localhost:<host port>/...` from the host. smolvm's bundled crane auto-falls-back to HTTP for localhost addresses, so no insecure-registries config is needed on that side. 6. Tear down everything; reap the tarball (registries hold the same bytes, no need to keep both around). Net effect: the docker daemon never does an HTTP/HTTPS-policy decision on our behalf. `docker push` is gone from the prepare path; `docker save`, `docker network create`, `docker run` (for registry + crane) replace it. Tested end-to-end on Docker Desktop / macOS: `_ensure_smolmachine ("claude-bottle:latest")` produces a 204MB `.smolmachine.smolmachine` artifact. Adds: - backend/docker/util.py:save() — thin docker save wrapper. - local_registry.crane_push_tarball() — one-shot crane run on the registry's network. - CRANE_IMAGE constant pinned by digest (gcr.io/go-containerregistry/crane@sha256:0ae17ecb...). Removes: - backend/docker/util.py:tag() / push() — unused without daemon push. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Tests
Plain-Python test suite using stdlib unittest. No external
dependencies. Unit tests run anywhere Python 3 is present; integration
tests need Docker and skip cleanly otherwise.
Layout
tests/
fixtures.py # JSON manifest builders (shared)
_docker.py # docker-availability skip helper (shared)
unit/
test_pipelock_classify.py
test_pipelock_allowlist.py
test_pipelock_yaml.py
test_manifest_runtime.py
integration/
test_pipelock_sidecar_smoke.py
test_dry_run_plan.py
test_orphan_cleanup.py
canaries/
test_pipelock_image.py # opt-in; see below
Classification falls out of the directory — no hand-maintained list to keep in sync.
Running
python -m unittest discover -t . -s tests/unit -v # unit only
python -m unittest discover -t . -s tests/integration -v # integration only
python -m unittest discover -t . -s tests -v # both (recursive)
python -m unittest tests.unit.test_pipelock_yaml # one file
Discovery is invoked with -t . (top-level dir = repo root) so the
claude_bottle package on sys.path resolves correctly.
What the integration tests cover
test_dry_run_plan.py—cli.py start --dry-run --format=jsonemits a structured plan that contains the resolved egress allowlist and the bottle's runtime, and creates zero Docker resources.test_orphan_cleanup.py—network_removeis idempotent against missing resources, so the EXIT trap can call it unconditionally.test_sidecar_bundle_image.py— builds Dockerfile.sidecars and probes that pipelock / gitleaks / mitmdump / supervise are all reachable inside the bundle.test_sidecar_bundle_compose.py— end-to-end compose-up of an agent + bundle pair; verifies the agent reaches the bundle via the legacy network aliases.
Canaries
tests/canaries/ holds upstream-regression checks (e.g. the pinned
pipelock digest's binary still runs). These are gated on
CLAUDE_BOTTLE_RUN_CANARIES=1 and not part of the per-push suite.
They're invoked by the scheduled canaries workflow.
CLAUDE_BOTTLE_RUN_CANARIES=1 python -m unittest discover -t . -s tests/canaries -v
What's NOT covered
claude_bottle/ssh.pyend-to-end (would need a fake SSH host inside the container).- A live SSH-through-pipelock tunnel against a real Tailscale-style IP.
- DLP false-positive measurements.
- TLS handling / cert pinning behavior.
Adding a test
- Pick the directory:
tests/unit/for a pure unit test,tests/integration/for one that needs Docker. - Filename:
test_<topic>.py. - Boilerplate:
import unittest from claude_bottle.<module> import <symbol> class TestThing(unittest.TestCase): def test_x(self): ... if __name__ == "__main__": unittest.main() - For Docker-dependent tests, decorate the class with
@skip_unless_docker()fromtests._docker.