feat(sandbox): the confinement seam and the per-platform native runner chains

ctx.sandbox (dsh-sandbox): confine(argv, policy) returns the argv to spawn
instead — wrapped so the process and its children run confined — plus the
enforcement completeness and the backend denial/runner-failure dialects;
no usable backend throws the fail-closed SANDBOX_UNAVAILABLE. Policy rides
per call. dsh-sandbox-local selects by platform and caches the verdict:
multi-candidate chains probe FUNCTIONALLY in preference order (Linux:
bwrap → the registry-installed node-addon-landlock-run launcher), a sole
candidate is selected unprobed (darwin: sandbox-exec/Seatbelt) and fails
closed at execution via runnerFailureSignatures; win32 is a reserved empty
chain. Profile parity is honest per backend (documented temp-area and ABI
differences; enforcement full|partial is a structured result fact).

CI: the sandbox-e2e matrix proves real-kernel confinement per rung (bwrap,
Landlock per architecture through the registry-installed launcher,
Seatbelt), failing on a silent all-skip; the packed-install rehearsal
installs the launcher family from the registry and asserts the binary
executable apart from kernel enforcement.
This commit is contained in:
kingwl
2026-07-09 15:42:37 +08:00
parent 80d8726601
commit 7b8c3a9b40
30 changed files with 2007 additions and 3 deletions

124
.github/workflows/sandbox.yml vendored Normal file
View File

@@ -0,0 +1,124 @@
# Sandbox CI: the keyless real-kernel confinement proofs. A separate workflow
# from ci.yml because the axis is different — these jobs fan out over
# OS×runner (kernel capabilities), not node versions. The Landlock launcher
# arrives from the registry with `pnpm install` (the npm package family
# `node-addon-landlock-run`, built and released from its own repository), so
# these legs exercise the true consumer path — nothing is compiled here.
name: Sandbox
on:
push:
branches: [main, master]
pull_request:
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: true
permissions:
contents: read
jobs:
# Keyless real-kernel sandbox proofs (sandbox RFC § Testing): each ladder
# rung is only provable on a host where it enforces, so this job fans out
# an OS×runner matrix — bwrap and Landlock on Linux (separate legs: the
# Landlock files force the bwrap rung off, so each leg proves exactly one
# rung; Landlock twice, once per architecture, each confining through the
# registry-installed launcher), Seatbelt on macOS (sandbox-exec ships with
# the OS). One node
# version only: kernel confinement does not vary by node, and ci.yml's
# node matrix already covers the node axis.
#
# The e2e files self-skip where their runner is absent, so a leg that lost
# its runner (no bwrap, kernel without Landlock, macOS without
# sandbox-exec) would otherwise pass as a false green — the same trap
# e2e.yml's key preflight guards against. Each leg therefore asserts BOTH
# its platform files actually ran: `Test Files 2 passed (2)`, no skips.
sandbox-e2e:
strategy:
fail-fast: false
matrix:
include:
- os: ubuntu-latest
runner: bwrap
- os: ubuntu-24.04
runner: landlock
- os: ubuntu-24.04-arm
runner: landlock
- os: macos-latest
runner: seatbelt
name: sandbox e2e (${{ matrix.runner }}, ${{ matrix.os }})
runs-on: ${{ matrix.os }}
timeout-minutes: 20
steps:
- uses: actions/checkout@v6
- uses: actions/setup-node@v6
with:
node-version: 24
- name: Enable corepack (pnpm)
run: corepack enable
- name: Install (immutable)
run: pnpm install --frozen-lockfile
# The bwrap rung needs bubblewrap on PATH and unprivileged user
# namespaces. Ubuntu 24.04 gates the latter behind an AppArmor knob;
# lift it best-effort — on images where the knob is absent the
# functional probe (and the run-guard below) is the arbiter anyway.
- name: Install bubblewrap (unrestrict userns)
if: matrix.runner == 'bwrap'
run: |
sudo apt-get update -q
sudo apt-get install -yq bubblewrap
sudo sysctl -w kernel.apparmor_restrict_unprivileged_userns=0 \
|| echo "apparmor userns knob absent — the functional probe decides"
# The unit suite runs on ubuntu in `checks`; this is the one darwin leg
# in the workflow, so run it here too — the platform-dependent unit
# expectations (Seatbelt path canonicalization: /tmp IS /private/tmp)
# take their darwin branch only on this runner.
- name: Unit tests (darwin parity)
if: matrix.runner == 'seatbelt'
run: pnpm run test
- name: Sandbox e2e (real kernel confinement, world-verified)
# NO_COLOR: vitest force-enables ANSI color under GITHUB_ACTIONS even
# without a TTY, which would thread escape codes through the summary
# line the run-guard greps.
env:
NO_COLOR: 1
run: |
set -u +e -o pipefail
out=$(pnpm exec vitest run --config vitest.e2e.config.ts \
packages/sandbox/sandbox-local/tests/${{ matrix.runner }}.e2e.ts \
packages/bash/bash-sandbox/tests/${{ matrix.runner }}.e2e.ts 2>&1); status=$?
echo "$out"
[ "$status" -eq 0 ]
# Both platform files must have RUN — a self-skip (runner missing on
# the very platform that exists to prove it) is a failure, not a pass.
echo "$out" | grep -qE 'Test Files[[:space:]]+2 passed \(2\)'
# Publish-path rehearsal, Landlock legs only (the pack gates need built
# lib/). The e2e packs the workspace closure, installs the tarballs
# into a throwaway consumer — npm pulling `node-addon-landlock-run`
# and its platform package from the registry, the true consumer path —
# and confines through the INSTALLED launcher, asserting it executable
# apart (a mode-stripped binary must not masquerade as a non-enforcing
# kernel). Same no-silent-skip guard as above.
- name: Build packages (lib/ for the pack rehearsal)
if: matrix.runner == 'landlock'
run: pnpm run build
- name: Packed-distribution e2e (pack → install → confine)
if: matrix.runner == 'landlock'
env:
NO_COLOR: 1
run: |
set -u +e -o pipefail
out=$(pnpm exec vitest run --config vitest.e2e.config.ts \
packages/sandbox/sandbox-local/tests/packed-install.e2e.ts 2>&1); status=$?
echo "$out"
[ "$status" -eq 0 ]
echo "$out" | grep -qE 'Test Files[[:space:]]+1 passed \(1\)'