748 Commits
Author SHA1 Message Date
Mesut Oezdil cd70249629 ci: pin azure/setup-helm and helm/kind-action to commit SHAs (#1544)
* ci: pin azure/setup-helm and helm/kind-action to commit SHAs

* chore(python): add py.typed marker for PEP 561 compliance

* ci: use full semver in pinned action version comments

Signed-off-by: mesutoezdil <mesudozdil@gmail.com>

---------

Signed-off-by: mesutoezdil <mesudozdil@gmail.com>
v0.0.49
2026-05-25 14:03:32 -07:00
Piotr Mlocek 286ce7c61a ci(release): skip python rpm in gateway smoke test (#1559)
Signed-off-by: Piotr Mlocek <pmlocek@nvidia.com>
v0.0.48
2026-05-25 13:41:04 -07:00
Mesut Oezdil 863d2a2ea9 chore(helm): add missing SPDX header to gateway-config template (#1545)
* chore(helm): add missing SPDX header to gateway-config template

* chore(scripts): remove helm templates from license header exclusions

The bypass had no known rationale. Removing it ensures the header
script covers deploy/helm/openshell/templates uniformly going forward.

Signed-off-by: mesutoezdil <mesudozdil@gmail.com>

---------

Signed-off-by: mesutoezdil <mesudozdil@gmail.com>
2026-05-25 12:28:47 -07:00
dependabot[bot] 5c3a1f7f3a chore(deps): bump docker/login-action from 4.1.0 to 4.2.0 (#1554)
Bumps [docker/login-action](https://github.com/docker/login-action) from 4.1.0 to 4.2.0.
- [Release notes](https://github.com/docker/login-action/releases)
- [Commits](https://github.com/docker/login-action/compare/4907a6ddec9925e35a0a9e82d7399ccc52663121...650006c6eb7dba73a995cc03b0b2d7f5ca915bee)

---
updated-dependencies:
- dependency-name: docker/login-action
  dependency-version: 4.2.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-05-25 11:53:51 -07:00
Piotr Mlocek c8d405cc37 ci(release): smoke test rpm artifacts on fedora (#1558)
Signed-off-by: Piotr Mlocek <pmlocek@nvidia.com>
2026-05-25 11:20:24 -07:00
Adel Zaalouk f0f17bf42a fix(cli): propagate --gateway-insecure to OIDC auth flows (#1535)
Thread the gateway_insecure flag through gateway_add(), gateway_login(),
and all OIDC HTTP clients so that --gateway-insecure and
OPENSHELL_GATEWAY_INSECURE apply to OIDC discovery, token exchange, and
token refresh requests.

Previously, the flag only affected gRPC connections to the gateway. OIDC
HTTP clients (reqwest::get and http_client) always verified TLS
certificates, causing gateway registration and login to fail when the
OIDC issuer used a self-signed certificate (common on OpenShift with
edge-terminated routes).

Fixes #1534

Signed-off-by: Adel Zaalouk <azaalouk@redhat.com>
2026-05-22 13:40:12 -07:00
Piotr Mlocek fbd580b2af ci: install cargo-zigbuild from release binaries (#1533) 2026-05-22 12:40:09 -07:00
Taylor Mutch 7d38aa8b8e fix(homebrew): repair local driver bootstrap state (#1527)
* fix(homebrew): repair local driver bootstrap state

* fix(bootstrap): satisfy default SAN doc lint
2026-05-22 12:12:39 -07:00
Piotr Mlocek 0dc08a185e fix(release): build host Linux binaries with glibc floor (#1490) 2026-05-22 12:12:03 -07:00
Piotr Mlocek 521eccd47e ci: seed shared Rust caches from main (#1530) 2026-05-22 12:11:10 -07:00
Drew Newberry 603b3e27fa docs: update NemoClaw/OpenClaw references (#1529) 2026-05-22 10:31:59 -07:00
John T. Myers 0cef26521a feat(providers): derive discovery from profiles (#1503)
* feat(providers): derive discovery from profiles

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* fix(providers): keep v2 discovery profile-only

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): update providers v2 behavior

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* fix(providers): make github profile read-only

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

---------

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>
2026-05-22 08:54:15 -07:00
mjamivandJohn Myers 686b24da29 fix(cli): add json output for policy get (#1410)
* fix(cli): add json output for policy get

* test(cli): cover policy get full json output

* fix(cli): address policy get json clippy

---------

Co-authored-by: John Myers <9696606+johntmyers@users.noreply.github.com>
v0.0.47
2026-05-22 07:39:27 -07:00
Taylor Mutch 48333e5e75 ci(canary): keep helm jwt secret generation enabled (#1521)
Signed-off-by: Taylor Mutch <taylormutch@gmail.com>
2026-05-22 07:21:57 -07:00
Evan Lezar 18988bd3b1 docs(rfc): add sandbox resource requirements proposal (#1360)
* docs(rfc): add sandbox resource requirements proposal

Signed-off-by: Evan Lezar <elezar@nvidia.com>

* docs(rfc): finalize sandbox resource requirements

---------

Signed-off-by: Evan Lezar <elezar@nvidia.com>
2026-05-22 07:11:11 -07:00
Taylor Mutch 57b71c68f2 ci(e2e): load single-arch images into kind (#1518)
Signed-off-by: Taylor Mutch <taylormutch@gmail.com>
2026-05-22 06:06:58 -07:00
Taylor Mutch 68d428055e fix(docker): use host-gateway callbacks on macOS (#1516) 2026-05-22 06:02:38 -07:00
John T. Myers c5c3f03ce7 docs(sandboxes): add policy advisor guide (#1480)
Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>
2026-05-21 18:06:42 -07:00
Taylor Mutch a3b16c18ab feat(auth): per-sandbox authentication to gateway (#1404) 2026-05-21 17:58:55 -07:00
Taylor Mutch 52389370b4 ci: deduplicate e2e workflows (#1512)
Signed-off-by: Taylor Mutch <taylormutch@gmail.com>
2026-05-21 17:44:05 -07:00
Adam Miller f5b0ad7134 fix(packaging): add upgrade migration docs and podman socket retry (#1507)
After #1415 ships, users upgrading from previous releases need guidance
on the gateway.env deprecation, port/bind/database path changes, and
the podman.socket restart requirement.

- docs(rpm): add 'Migrating from gateway.env' section to TROUBLESHOOTING
  covering backward compatibility, env-to-TOML key mapping, and three
  breaking changes (default port 8080->17670, bind address 0.0.0.0->127.0.0.1,
  database path move). Add podman.socket restart step to upgrade procedure.
- docs(rpm): add upgrade callout to CONFIGURATION.md pointing at migration
  section.
- fix(podman): retry PodmanComputeDriver ping up to 5 times with 2s delay
  to tolerate transient socket unavailability after package upgrades.
  The systemd unit uses Wants=podman.socket (not Requires) so the gateway
  can start while the socket is briefly re-activating after an RPM upgrade
  changes its unit file on disk.
- chore(rpm): update EnvironmentFile comment in RPM spec to explain
  backward-compatibility intent.

Signed-off-by: Adam Miller <admiller@redhat.com>
2026-05-21 15:57:13 -07:00
Colin Walters e7f965a988 refactor(sandbox,driver-vm): Start moving to rustix (esp over libc unsafe) (#1505)
In the Rust ecosystem there's largely three ways to do system calls:

- raw libc
- nix
- rustix

Of the three, libc is almost all `unsafe` and really 95% of use
cases should be either nix or rustix. nix is the original one,
but after having looked at the code of both, I think rustix
is just better designed and organized. It's also reached 1.0,
whereas nix is still making semver-breaking changes (in fact
we're behind here in this project).

Now in practice, we have both *transitively* in the depchain
already, and that's true for quite a lot of projects.

But I think rustix is better, so let's add rustix as
a workspace dependency (process feature) and migrate
a few use cases to it - it's especially better than the raw
libc which is suprisingly widespread.

If we agree to do this, then many other calls can be ported.

Signed-off-by: Colin Walters <walters@verbum.org>
2026-05-21 15:20:21 -05:00
Russell Bryant af753748d3 fix(server): respect OPENSHELL_PODMAN_SOCKET env var in embedded driver (#1483)
The env var was only wired up via clap in the standalone
openshell-driver-podman binary. When the Podman driver runs embedded
in the gateway, config came exclusively from TOML deserialization and
the env var was never consulted. Apply it as a post-deserialization
override, matching the existing OPENSHELL_K8S_WORKSPACE_DEFAULT_STORAGE_SIZE
pattern.

Closes #1446
2026-05-21 15:03:48 -05:00
Piotr Mlocek f8e3f9b3cd fix(ci): resolve mirror gate statuses for fork PRs (#1504)
Signed-off-by: Piotr Mlocek <pmlocek@nvidia.com>
2026-05-21 11:25:07 -07:00
Eric Curtin 5620c8b229 refactor: deduplicate repeated patterns across crates (#1499)
Remove ~280 lines of duplicated code across 30 files in 5 areas:

- centered_rect: consolidate 5 identical TUI layout helpers into a
  single pub fn in openshell-tui/src/ui/mod.rs
- server test helpers: replace ~100 inline Store::connect() calls
  with local test_store() helpers; deduplicate test_server_state()
  in grpc/service.rs to use the shared test_support version
- rogue PKI: extract 20-line rogue CA+client cert generation block
  (duplicated in two integration tests) into generate_rogue_pki()
  in tests/common/mod.rs
- provider tests: replace 8 identical 28-line test modules with a
  single macro_rules! test_discovers_env_credential! invocation
- label constants: centralize openshell.ai/ container label keys
  in openshell-core::driver_utils; update Docker and Kubernetes
  drivers to import from there instead of redefining them locally
2026-05-21 10:50:49 -07:00
Florent BENOIT 9e8610f6ea feat(cli): add JSON/YAML output format to gateway list (#1500)
Add -o/--output flag to `openshell gateway list` matching the existing
sandbox list pattern, enabling machine-readable output for scripting.

Signed-off-by: Florent Benoit <fbenoit@redhat.com>
2026-05-21 10:39:20 -07:00
mjamivandJohn Myers 528fb29147 fix(sandbox): allow first-label L7 host wildcards (#1304)
* fix(sandbox): allow first-label L7 host wildcards

* docs(sandbox): document L7 host wildcard contract + add OPA runtime tests

- Add Host Wildcards section to architecture/security-policy.md
  describing accepted (first-label *, **, intra-label *-X) and
  rejected (bare, TLD, non-first-label, recursive-in-label) forms,
  and noting that wildcards never cross '.' boundaries.
- Expand the policy-schema.mdx 'host' field description to reflect
  the same contract instead of only mentioning '*.example.com'.
- Add OPA runtime tests asserting '*-aiplatform.googleapis.com'
  matches 'us-central1-aiplatform.googleapis.com' and does not match
  'us-central1.aiplatform.googleapis.com' (cross-dot boundary). Locks
  validator/runtime alignment for intra-label wildcards.

* chore: update mise lockfile

* test(server): tolerate serialized inference upserts

---------

Co-authored-by: John Myers <9696606+johntmyers@users.noreply.github.com>
2026-05-21 09:32:37 -07:00
Derek Carr 2d9e5326f8 fix(sandbox): skip fork-exec socket ambiguity test on SELinux-enforcing hosts (#1449)
Exec'ing /bin/sleep (SELinux label bin_t) from a user_home_t test binary
causes /proc/<pid>/exe readlink to return ENOENT on SELinux-enforcing
hosts due to the cross-domain boundary. Skip the test at runtime when
getenforce reports Enforcing.

Also adds a ChildGuard drop guard for safe child cleanup on panic and
increases the exec-detection deadline from 2s to 5s.

Signed-off-by: Derek Carr <decarr@redhat.com>
2026-05-21 09:13:57 -07:00
Eric Curtin e3f009fb9a refactor(server): extract shared relay-await and sandbox-scan helpers (#1495) v0.0.46 2026-05-21 07:23:36 -07:00
Florent BENOIT b93a3d8661 fix(scripts): use portable lowercase in normalize_bool for Bash 3.2 (#1493) 2026-05-21 07:15:07 -07:00
Seth Jennings c143c81fbf fix(cli): add auth and TLS support to completion client (#1489) 2026-05-21 06:52:23 -07:00
Drew NewberryandTaylor Mutch 77e6c7a12a test(server): cover service endpoint plaintext security (#1352)
* test(server): cover service endpoint plaintext security

* test(server): align tls test with from_files Option<&Path> signature

TlsAcceptor::from_files now accepts the client CA path as Option<&Path>
(per the require_client_auth refactor on main). Wrap the helper's CA
path in Some(...) so the new plaintext-service-http tests compile after
rebasing onto current main.

---------

Co-authored-by: Taylor Mutch <taylormutch@gmail.com>
2026-05-20 14:44:16 -07:00
Mesut Oezdil 2b13bfa47b fix(ci): eliminate image-tag race between concurrent workflows (#1413)
- Add publish-manifest input to docker-build.yml (default true); single-arch
  branch callers set it false so the merge job is skipped and the shared
  bare :SHA tag in GHCR is never written by branch workflows
- branch-kubernetes-e2e: retag :SHA-amd64 to :SHA before kind load so
  Helm's image.tag matches what is loaded in kind containerd
- branch-e2e: pass image-tag as :SHA-arm64 to e2e-test so the arch-specific
  GHCR tag is used directly without depending on the bare tag
- bare :SHA in GHCR is now written only by test-gpu.yml (multi-arch build),
  eliminating the last-writer-wins race across concurrent workflows
2026-05-20 14:06:25 -07:00
Adel Zaalouk bdaa08fbb0 fix(server): add ConnectSupervisor and RelayStream to SANDBOX_METHODS (#1475) 2026-05-20 10:01:56 -07:00
dependabot[bot] 3cde651834 chore(deps): bump azure/setup-helm from 4 to 5 (#1468)
Bumps [azure/setup-helm](https://github.com/azure/setup-helm) from 4 to 5.
- [Release notes](https://github.com/azure/setup-helm/releases)
- [Changelog](https://github.com/Azure/setup-helm/blob/main/CHANGELOG.md)
- [Commits](https://github.com/azure/setup-helm/compare/v4...v5)

---
updated-dependencies:
- dependency-name: azure/setup-helm
  dependency-version: '5'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-05-20 09:35:48 -07:00
Taylor Mutch c600b11ff2 docs(agents): add release canary testing skill (#1440)
* ci(canary): add kind-based helm chart smoke test

Signed-off-by: Taylor Mutch <taylormutch@gmail.com>

* docs(agents): add release canary testing skill

Signed-off-by: Taylor Mutch <taylormutch@gmail.com>

---------

Signed-off-by: Taylor Mutch <taylormutch@gmail.com>
2026-05-20 09:34:57 -07:00
Eric Curtin b332ffdf7d refactor: deduplicate shared driver and provider constants (#1474)
Move default_sandbox_image() and SUPERVISOR_IMAGE_BINARY_PATH into
openshell-core so all compute drivers share a single authoritative
definition. Eliminates four identical copies of default_sandbox_image()
and two identical copies of SUPERVISOR_IMAGE_BINARY_PATH.

Implement ProviderPlugin for ProviderDiscoverySpec to remove the
repeated three-method delegation boilerplate from eight simple
provider modules. Each now declares only a SPEC constant; the blanket
impl handles id(), discover_existing(), and credential_env_vars().
2026-05-20 09:24:57 -07:00
John T. Myers 14c5329f91 docs(providers): add Providers v2 guide (#1442)
* docs(providers): add Providers v2 guide

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): clarify built-in profile location

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): remove refresh-backed profile bullet

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): clarify persisted query registry key

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): simplify binary schema example

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): add Microsoft Graph refresh tutorial

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): link Microsoft Graph refresh tutorial

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

* docs(providers): announce providers v2

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>

---------

Signed-off-by: John Myers <9696606+johntmyers@users.noreply.github.com>
2026-05-20 08:57:24 -07:00
Evan Lezar be6ac9e61c docs(agents): add Docker GPU CDI debug hints (#1448)
Signed-off-by: Evan Lezar <elezar@nvidia.com>
v0.0.45
2026-05-20 14:43:54 +02:00
Evan Lezar cade0bb0ac test(e2e): default GPU probe image (#1450)
Signed-off-by: Evan Lezar <elezar@nvidia.com>
2026-05-20 13:25:09 +02:00
Taylor Mutch 2a065a5ff2 ci(canary): add kind-based helm chart smoke test (#1336) 2026-05-19 17:25:57 -07:00
Derek Carr 3c8739372f feat(agents): add LSM compatibility checks to review and spike skills (#1451)
Teach the principal-engineer-reviewer agent, build-from-issue skill, and
create-spike skill to assess Linux Security Module impact.  Code that
touches /proc filesystem access, process identity, or binary execution
can behave differently on SELinux-enforcing or AppArmor hosts — for
example, readlink on /proc/<pid>/exe returns ENOENT across SELinux
domain boundaries, breaking tests that fork+exec into system binaries.

Signed-off-by: Derek Carr <decarr@redhat.com>
2026-05-19 17:19:27 -07:00
Piotr Mlocek 3b53184971 test(persistence): make CAS conflict test deterministic (#1464) 2026-05-19 16:46:52 -07:00
Zygmunt Krynicki 0a8b35c31f fix(build): install binaries built in part build tree (#1462)
The earlier build only worked if the tree had pre-compiled binaries left
over from non-snap build.

Signed-off-by: Zygmunt Krynicki <zygmunt.krynicki@canonical.com>
2026-05-19 15:33:25 -07:00
dependabot[bot] 37ca269072 chore(deps): bump softprops/action-gh-release from 2.6.2 to 3.0.0 (#1458)
Bumps [softprops/action-gh-release](https://github.com/softprops/action-gh-release) from 2.6.2 to 3.0.0.
- [Release notes](https://github.com/softprops/action-gh-release/releases)
- [Changelog](https://github.com/softprops/action-gh-release/blob/master/CHANGELOG.md)
- [Commits](https://github.com/softprops/action-gh-release/compare/3bb12739c298aeb8a4eeaf626c5b8d85266b0e65...b4309332981a82ec1c5618f44dd2e27cc8bfbfda)

---
updated-dependencies:
- dependency-name: softprops/action-gh-release
  dependency-version: 3.0.0
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-05-19 15:17:39 -07:00
dependabot[bot] 2cef120dc1 chore(deps): bump actions/download-artifact from 4.3.0 to 8.0.1 (#1459)
Bumps [actions/download-artifact](https://github.com/actions/download-artifact) from 4.3.0 to 8.0.1.
- [Release notes](https://github.com/actions/download-artifact/releases)
- [Commits](https://github.com/actions/download-artifact/compare/d3f86a106a0bac45b974a628896c90dbdf5c8093...3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c)

---
updated-dependencies:
- dependency-name: actions/download-artifact
  dependency-version: 8.0.1
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2026-05-19 15:17:23 -07:00
Piotr Mlocek 2a5a44989a fix(ci): require PR checks to pass (#1461) 2026-05-19 14:46:23 -07:00
Eric Curtin 10af3e609f refactor: deduplicate shared test helpers (#1399)
Extract four categories of copy-pasted test code into single
canonical locations:

- POLICY_OBJECT_TYPE / DRAFT_CHUNK_OBJECT_TYPE constants that were
  defined identically in both persistence/sqlite.rs and
  persistence/postgres.rs are now owned by persistence/mod.rs.

- test_server_state() was duplicated across grpc/provider.rs,
  grpc/policy.rs, and grpc/sandbox.rs. A shared grpc::test_support
  module in grpc/mod.rs now owns the single implementation; each
  submodule imports it.

- The TestOpenShell stub (full OpenShell trait impl), install_rustls_provider(),
  PkiBundle, generate_pki(), and start_test_server() were copied across
  up to five openshell-server integration test files. A new
  tests/common/mod.rs module owns them; each test file uses mod common.

- StubResponse, unique_socket_path(), and spawn_podman_stub() were
  duplicated between openshell-driver-podman/src/driver.rs and
  openshell-driver-podman/src/grpc.rs. A new src/test_utils.rs
  (cfg(test)-gated) owns the shared helpers.
2026-05-19 12:26:12 -07:00
Seth Jennings c527341d8a feat(k8s): make default workspace PVC storage size configurable (#1436)
The Kubernetes driver hardcoded the workspace PVC size to 2Gi. Add a
workspace_default_storage_size field to KubernetesComputeConfig so
operators can tune it via TOML config, Helm values, or the
OPENSHELL_K8S_WORKSPACE_DEFAULT_STORAGE_SIZE environment variable.
2026-05-19 11:42:16 -07:00
John T. Myers d255cdd9c9 feat(providers): add credential refresh foundation (#1349)
* feat(providers): add credential refresh foundation

* feat(providers): mint oauth refresh token credentials

* fix(providers): allow delegated refresh bootstrap

* fix(providers): guard refresh credential modes

* fix(providers): refine refresh lifecycle UX

* fix(cli): restore provider test gateway response import

* fix(server): clean auth endpoint test qualifications

* fix(providers): tighten refresh authorization and collisions

* chore(providers): trim bundled v2 profiles

* fix(providers): resolve refresh rebase fallout

* fix(cli): accept rfc3339 credential expiry

* test(providers): update inferred claude provider type

* test(providers): avoid removed outlook default profile

* test(providers): isolate attach limit fixtures
v0.0.44
2026-05-19 07:42:10 -07:00