hermes-agent/tests/docker/test_immutable_install.py
Teknium 39975613b1
test: prune wave 2 + speed fixes — 28,106 → 19,757 test functions, suite wall 315s → 294s
Second, deeper pass over tools/gateway/hermes_cli plus first pass over
the trees wave 1 missed (acp, acp_adapter, skills, computer_use, docker,
dashboard, conformance, monitoring, secret_sources, hermes_state,
providers). Same rubric as wave 1 (AGENTS.md test policy); security,
alternation/caching invariants, issue-number regressions, and E2E kept.

Real test-quality fixes found and rooted out along the way:
- tests/tools/test_command_guards.py made real auxiliary-LLM HTTPS calls
  (DEFAULT_CONFIG smart-approval leaked in) — pinned approval
  mode=manual via autouse fixture: 17.4s → 0.4s.
- test_model_switch_custom_providers.py / test_user_providers_model_switch.py
  silently probed live provider catalogs (~2s/test) — stubbed
  cached_provider_model_ids/provider_model_ids/fetch_api_models.
- test_telegram_noise_filter.py: 15-platform copy-paste matrix over
  shared gateway.run logic → 3 representative platforms (55s → 3.9s).
- test_gateway_shutdown.py: stop()'s 5s interrupt-deadline loop spun on
  MagicMock agents — interrupt.side_effect now clears _running_agents
  (22s → 1.0s).
- test_gateway_inactivity_timeout.py poll-harness timings shrunk 3-5x
  (24s → 1.1s); test_mcp_stability.py backoff/SIGTERM-grace sleeps
  patched (15.4s → 2.5s); test_async_delegation.py negative-drain wait
  5s → 0.5s.
- test_telegram_init_deadline.py: loop-block margin restored to 1.0s
  with rationale comment — the watchdog-dump assertion needs the loop
  blocked well past deadline+grace under parallel load (flaked once in
  the 40-worker verification run at a 0.2s margin).

Verification: full hermetic suite via scripts/run_tests.sh —
2,438 files, 21,718 tests passed, 0 failed, 293.9s wall.
Suite totals vs original baseline: 46,820 → 19,757 test functions
(−57.8%), wall 583.5s → 293.9s (−50%), subprocess CPU 13,564s → 11,623s.
2026-07-29 13:39:40 -07:00

110 lines
3.6 KiB
Python

"""Runtime smoke tests for Docker immutable install tree and install-method stamp.
Build the real image and verify at runtime:
1. /opt/hermes is not writable by the hermes user (immutable install tree)
2. PYTHONDONTWRITEBYTECODE and HERMES_DISABLE_LAZY_INSTALLS are set
3. /opt/hermes/.install_method contains "docker" (code-scoped stamp)
4. $HERMES_HOME/.install_method is NOT stamped as "docker" by stage2
5. A stale "docker" stamp in $HERMES_HOME is healed (removed) on boot
"""
from __future__ import annotations
from tests.docker.conftest import (
docker_exec,
docker_exec_sh,
restart_container,
start_container,
)
def test_install_tree_not_writable_by_hermes(
built_image: str, container_name: str,
) -> None:
"""The hermes user must not be able to modify /opt/hermes.
The install tree (source, venv, TUI bundle, node_modules) must remain
root-owned and non-writable so an agent session cannot self-modify
the installation and brick the gateway.
"""
start_container(built_image, container_name)
r = docker_exec_sh(
container_name,
# Try to create a file under /opt/hermes as the hermes user
"touch /opt/hermes/test_write 2>&1 && "
"echo WRITE_SUCCEEDED || echo WRITE_FAILED",
timeout=10,
)
assert "WRITE_FAILED" in r.stdout, (
f"hermes user can write to /opt/hermes (install tree not immutable): "
f"{r.stdout}"
)
# Also check a key subdirectory
r = docker_exec_sh(
container_name,
"touch /opt/hermes/.venv/test_write 2>&1 && "
"echo WRITE_SUCCEEDED || echo WRITE_FAILED",
timeout=10,
)
assert "WRITE_FAILED" in r.stdout, (
f"hermes user can write to /opt/hermes/.venv: {r.stdout}"
)
def test_hermes_disable_lazy_installs_and_dont_write_bytecode(
built_image: str, container_name: str,
) -> None:
"""The container must set PYTHONDONTWRITEBYTECODE and
HERMES_DISABLE_LAZY_INSTALLS=1 so no .pyc files are written to the
immutable install tree and no lazy installs attempt to modify it."""
start_container(built_image, container_name)
r = docker_exec_sh(
container_name,
'test "$PYTHONDONTWRITEBYTECODE" = "1" && '
'test "$HERMES_DISABLE_LAZY_INSTALLS" = "1" && '
'echo ENV_OK || echo ENV_MISSING',
timeout=10,
)
assert "ENV_OK" in r.stdout, (
f"expected PYTHONDONTWRITEBYTECODE=1 and "
f"HERMES_DISABLE_LAZY_INSTALLS=1, got: {r.stdout} stderr={r.stderr}"
)
def test_stale_docker_stamp_in_home_is_healed_on_boot(
built_image: str, container_name: str,
) -> None:
"""A stale 'docker' stamp left in $HERMES_HOME by an older image
must be removed on boot so shared homes self-heal."""
# Start container, write a stale stamp
start_container(built_image, container_name)
# Write a stale 'docker' stamp as root
docker_exec(
container_name, "sh", "-c",
"printf 'docker\\n' > /opt/data/.install_method",
user="root", timeout=5,
)
# Verify it exists
r = docker_exec_sh(container_name, "cat /opt/data/.install_method", timeout=5)
assert r.stdout.strip() == "docker"
# Restart - stage2 should heal it
restart_container(container_name)
# The stale stamp must be gone
r = docker_exec_sh(
container_name,
"test -f /opt/data/.install_method && "
"cat /opt/data/.install_method || echo HEALED",
timeout=10,
)
assert "HEALED" in r.stdout or r.stdout.strip() != "docker", (
f"stale 'docker' stamp in $HERMES_HOME was not healed on boot: "
f"{r.stdout}"
)