atlas-iac/testing/tests/test_hermes_agent_runtime_patches.py

436 lines
16 KiB
Python
Raw Normal View History

"""Pinned operator tooling and runtime patch contracts."""
from __future__ import annotations
import os
import subprocess
from testing.tests.test_hermes_cli_support import (
HERMES,
Path,
SCRIPTS,
_agent_deployment,
auth_patch,
client_config,
codex_runtime_patch,
json,
pytest,
ttyd_patch,
tui_gateway_patch,
yaml,
)
def test_owner_agent_installs_the_pinned_operator_toolchain():
script = (SCRIPTS / "install_agent_tools.sh").read_text()
for value in [
"flux",
"helm",
"kustomize",
"jq",
"yq",
"gh",
"vault",
"sops",
"age",
"age-keygen",
"k9s",
"terraform",
"go",
"gofmt",
]:
assert value in script
hermes(agent): make runtime tooling install architecture-aware The hermes-agent installs its CLI toolchain at runtime into the shared /opt/data/tools Longhorn volume, but every download hardcoded arm64. On the amd64 node titan-22 that left configure-agent-clients failing with "Missing optional dependency @openai/codex-linux-x64" and the operator toolchain fetching arm64 binaries, so the pod churned. Detect the running node's arch (uname -m; fail closed on anything but aarch64/x86_64) and resolve every asset per-arch: - install-agent-tools init script (agent-deployment.yaml): ttyd and kubectl download the arch-correct asset with the arch-correct sha256 (real ttyd 1.7.7 x86_64 and kubectl v1.33.3 amd64 checksums added; the arm64 ones kept). The npm CLI stamp is now arch-specific (.cli-versions-<vers>-${arch}) so a fresh arch re-runs npm install and pulls its own native optional deps; npm keeps both arches' packages. - install_agent_tools.sh: flux/helm/kustomize/jq/yq/gh/vault/sops/age/ k9s/terraform/go URLs, tarball subdirs (helm linux-${arch}, gh dir), and checksums are all arch-resolved with both arches pinned. Stamps and the Go tree are arch-specific, and an active-arch marker forces a republish of the single-arch ${bin} binaries when the pod moves between arches on the shared volume. Single fetch/verify helper kept. Tests updated to assert the arch-aware form (both arches' Go checksums, ${dl_arch} templating) instead of the arm64-only literal. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BvMSXH8VH2tMWXanb8SJdf
2026-08-25 17:37:57 -03:00
# The toolchain install resolves the running node's arch instead of a
# hardcoded one, so the Go URL is templated on ${dl_arch} and both arches'
# pinned Go checksums are present.
assert "go1.26.5.linux-${dl_arch}.tar.gz" in script
assert "dl_arch=arm64" in script
assert "dl_arch=amd64" in script
assert (
"fe4789e92b1f33358680864bbe8704289e7bb5fc207d80623c308935bd696d49"
in script
)
hermes(agent): make runtime tooling install architecture-aware The hermes-agent installs its CLI toolchain at runtime into the shared /opt/data/tools Longhorn volume, but every download hardcoded arm64. On the amd64 node titan-22 that left configure-agent-clients failing with "Missing optional dependency @openai/codex-linux-x64" and the operator toolchain fetching arm64 binaries, so the pod churned. Detect the running node's arch (uname -m; fail closed on anything but aarch64/x86_64) and resolve every asset per-arch: - install-agent-tools init script (agent-deployment.yaml): ttyd and kubectl download the arch-correct asset with the arch-correct sha256 (real ttyd 1.7.7 x86_64 and kubectl v1.33.3 amd64 checksums added; the arm64 ones kept). The npm CLI stamp is now arch-specific (.cli-versions-<vers>-${arch}) so a fresh arch re-runs npm install and pulls its own native optional deps; npm keeps both arches' packages. - install_agent_tools.sh: flux/helm/kustomize/jq/yq/gh/vault/sops/age/ k9s/terraform/go URLs, tarball subdirs (helm linux-${arch}, gh dir), and checksums are all arch-resolved with both arches pinned. Stamps and the Go tree are arch-specific, and an active-arch marker forces a republish of the single-arch ${bin} binaries when the pod moves between arches on the shared volume. Single fetch/verify helper kept. Tests updated to assert the arch-aware form (both arches' Go checksums, ${dl_arch} templating) instead of the arm64-only literal. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BvMSXH8VH2tMWXanb8SJdf
2026-08-25 17:37:57 -03:00
assert (
"5c2c3b16caefa1d968a94c1daca04a7ca301a496d9b086e17ad77bb81393f053"
in script
)
assert script.count("sha256sum -c -") == 1
deployment = _agent_deployment()
installer = next(
item
for item in deployment["spec"]["template"]["spec"]["initContainers"]
if item["name"] == "install-agent-tools"
)
assert "/bin/sh /opt/coordinator/install_agent_tools.sh" in installer["command"][2]
assert any(mount["name"] == "coordinator" for mount in installer["volumeMounts"])
install = installer["command"][2]
assert 'codex_version="$("${tools}/bin/codex" --version' in install
assert 'claude_version="$("${tools}/bin/claude" --version' in install
assert "@anthropic-ai" in install and "-name '.claude-code-*'" in install
assert "@openai" in install and "-name '.codex-*'" in install
assert "-mindepth 1 -maxdepth 1 -type d" in install
assert "[ ! -L \"${scope}\" ]" in install
init_config = next(
item
for item in deployment["spec"]["template"]["spec"]["initContainers"]
if item["name"] == "init-config"
)
init_command = init_config["command"][2]
assert "# Hermes managed operator PATH." in init_command
assert "/opt/coordinator:/opt/data/tools/bin:/opt/hermes/.venv/bin" in init_command
assert 'chmod 0644 "${profile_file}"' in init_command
def test_npm_reinstall_removes_only_known_interrupted_staging_directories(tmp_path):
"""A stale npm rename target is cleared before the pinned reinstall only."""
installer = next(
item
for item in _agent_deployment()["spec"]["template"]["spec"]["initContainers"]
if item["name"] == "install-agent-tools"
)
tools = tmp_path / "tools"
bin_dir = tools / "bin"
bin_dir.mkdir(parents=True)
for name, output in {
"ttyd": "ttyd 1.7.7\n",
"kubectl": '{"gitVersion":"v1.33.3"}\n',
"codex": "codex 0.154.0\n",
"claude": "2.1.226\n",
}.items():
path = bin_dir / name
path.write_text(f"#!/bin/sh\nprintf '%s' '{output}'\n", encoding="utf-8")
path.chmod(0o755)
for scope, stale in (("@anthropic-ai", ".claude-code-aborted"), ("@openai", ".codex-aborted")):
path = tools / "lib/node_modules" / scope / stale
path.mkdir(parents=True)
(path / "partial").write_text("incomplete", encoding="utf-8")
preserved = tools / "lib/node_modules/@anthropic-ai/keep-me"
preserved.mkdir(parents=True)
(preserved / "data").write_text("preserve", encoding="utf-8")
fake_npm = tmp_path / "fake-bin/npm"
fake_npm.parent.mkdir()
fake_npm.write_text(
"#!/bin/sh\n"
"test ! -e \"$HERMES_TEST_TOOLS/lib/node_modules/@anthropic-ai/.claude-code-aborted\"\n"
"test ! -e \"$HERMES_TEST_TOOLS/lib/node_modules/@openai/.codex-aborted\"\n",
encoding="utf-8",
)
fake_npm.chmod(0o755)
command = installer["command"][2]
command = command.replace("tools=/opt/data/tools", f"tools={tools}")
command = command.replace("/bin/sh /opt/coordinator/install_agent_tools.sh", ":")
command = command.replace('HOME=/opt/data/home git config --global user.name "Hermes Agent"', ":")
command = command.replace('HOME=/opt/data/home git config --global user.email "hermes@bstein.dev"', ":")
result = subprocess.run(
["sh", "-c", command],
text=True,
capture_output=True,
env={
**os.environ,
"PATH": f"{fake_npm.parent}:{os.environ['PATH']}",
"HERMES_TEST_TOOLS": str(tools),
},
)
assert result.returncode == 0, result.stderr
assert (preserved / "data").read_text(encoding="utf-8") == "preserve"
refactor(hermes): rename the owner agent host to worker.bstein.dev Introduce worker.bstein.dev as the canonical hostname for the owner-only Hermes coordinator, previously agent.hermes.bstein.dev. The rename is additive, matching the shape #38 restored for chat and triage. CoreDNS, both agent Ingresses and the hermes-sites certificate now serve BOTH names, so merging this cannot take away the endpoint the operator uses to reach the coordinator. Retiring agent.hermes.bstein.dev is a separate, separately scheduled change. No redirect middleware is added. What switches to the new host: - HERMES_DASHBOARD_PUBLIC_URL and the oauth2-proxy --redirect-url - the Keycloak agent proxy rootUrl - operator docs, skills, the ZAP baseline target and the triage monitor default What stays dual-homed until retirement: - CoreDNS hosts entry, both agent Ingress rules, certificate SANs - API_SERVER_CORS_ORIGINS (now a comma-separated pair) - the Keycloak redirect URIs, web origins and post-logout origins, so a rollback only needs the oauth2-proxy --redirect-url reverted and does not require re-running the ensure job The agent client passes its legacy origin through the optional fourth argument #38 added to ensure_proxy_client, so no second mechanism is introduced. The immutable ensure Job goes -11 -> -12 because #38 already consumed -11 and that run has completed; without a further bump this change would never be applied. Login on the new host fails until the -12 Job completes. Because the session and CSRF cookies use the __Host- prefix they are bound to one origin, so a fresh login must start on worker.bstein.dev and existing sessions do not carry over -- re-login is required after rollout. #38's public-host continuity test now covers the agent proxy's dual origins rather than asserting the agent surface was untouched by the rename. Knowledge catalogs and diagrams regenerated with `make knowledge`.
2026-08-21 10:29:46 +00:00
def test_owner_agent_uses_the_canonical_hostname():
paths = [
HERMES / "agent-configmap.yaml",
HERMES / "agent-deployment.yaml",
HERMES / "agent-ingress.yaml",
Path(__file__).parents[2] / "scripts/ops/hermes_triage_monitor.py",
]
for path in paths:
content = path.read_text()
assert "agent.bstein.dev" not in content
refactor(hermes): rename the owner agent host to worker.bstein.dev Introduce worker.bstein.dev as the canonical hostname for the owner-only Hermes coordinator, previously agent.hermes.bstein.dev. The rename is additive, matching the shape #38 restored for chat and triage. CoreDNS, both agent Ingresses and the hermes-sites certificate now serve BOTH names, so merging this cannot take away the endpoint the operator uses to reach the coordinator. Retiring agent.hermes.bstein.dev is a separate, separately scheduled change. No redirect middleware is added. What switches to the new host: - HERMES_DASHBOARD_PUBLIC_URL and the oauth2-proxy --redirect-url - the Keycloak agent proxy rootUrl - operator docs, skills, the ZAP baseline target and the triage monitor default What stays dual-homed until retirement: - CoreDNS hosts entry, both agent Ingress rules, certificate SANs - API_SERVER_CORS_ORIGINS (now a comma-separated pair) - the Keycloak redirect URIs, web origins and post-logout origins, so a rollback only needs the oauth2-proxy --redirect-url reverted and does not require re-running the ensure job The agent client passes its legacy origin through the optional fourth argument #38 added to ensure_proxy_client, so no second mechanism is introduced. The immutable ensure Job goes -11 -> -12 because #38 already consumed -11 and that run has completed; without a further bump this change would never be applied. Login on the new host fails until the -12 Job completes. Because the session and CSRF cookies use the __Host- prefix they are bound to one origin, so a fresh login must start on worker.bstein.dev and existing sessions do not carry over -- re-login is required after rollout. #38's public-host continuity test now covers the agent proxy's dual origins rather than asserting the agent surface was untouched by the rename. Knowledge catalogs and diagrams regenerated with `make knowledge`.
2026-08-21 10:29:46 +00:00
assert "worker.bstein.dev" in content
def test_legacy_owner_host_stays_served_until_it_is_retired_separately():
"""The rename must not cut off the host the operator reaches Hermes on.
``agent.hermes.bstein.dev`` keeps resolving, routing and validating TLS
until it is retired in its own change, so rolling the rename back never
needs more than reverting the oauth2-proxy callback.
"""
legacy = "agent.hermes.bstein.dev"
canonical = "worker.bstein.dev"
coredns = (
Path(__file__).parents[2] / "infrastructure/core/coredns-custom.yaml"
).read_text()
assert f"192.168.22.9 {legacy}" in coredns
assert f"192.168.22.9 {canonical}" in coredns
certificate = yaml.safe_load((HERMES / "agent-certificate.yaml").read_text())
assert {legacy, canonical} <= set(certificate["spec"]["dnsNames"])
ingresses = [
doc
for doc in yaml.safe_load_all((HERMES / "agent-ingress.yaml").read_text())
if doc
and doc.get("kind") == "Ingress"
and doc["metadata"]["name"].startswith("hermes-agent-")
]
assert len(ingresses) == 2
for ingress in ingresses:
assert {legacy, canonical} <= {rule["host"] for rule in ingress["spec"]["rules"]}
for tls in ingress["spec"]["tls"]:
assert {legacy, canonical} <= set(tls["hosts"])
def test_agent_reconnect_retains_complete_history_and_long_tool_budget():
configmap = yaml.safe_load((HERMES / "agent-configmap.yaml").read_text())
config = yaml.safe_load(configmap["data"]["config.yaml"])
display = config["display"]
assert display["resume_exchanges"] >= 10000
assert display["resume_max_user_chars"] >= 10000000
assert display["resume_max_assistant_chars"] >= 10000000
assert config["agent"]["max_turns"] == 180
assert config["delegation"]["max_iterations"] == 120
def test_auth_patch_honors_explicit_shared_store(tmp_path: Path):
source = tmp_path / "auth.py"
destination = tmp_path / "patched/auth.py"
source.write_text(
'from pathlib import Path\nimport os\n\ndef _auth_file_path() -> Path:\n path = get_hermes_home() / "auth.json"\n return path\n',
encoding="utf-8",
)
auth_patch.patch(source, destination)
content = destination.read_text()
assert 'os.environ.get("HERMES_AUTH_FILE"' in content
def test_auth_patch_fails_closed_on_upstream_drift(tmp_path: Path):
source = tmp_path / "auth.py"
source.write_text("def changed():\n pass\n", encoding="utf-8")
with pytest.raises(RuntimeError, match="context changed"):
auth_patch.patch(source, tmp_path / "patched.py")
def test_codex_runtime_patch_uses_cli_and_forwards_route(tmp_path: Path):
provider = tmp_path / "runtime_provider.py"
provider.write_text(codex_runtime_patch.PROVIDER_BEFORE, encoding="utf-8")
provider_out = tmp_path / "patched/runtime_provider.py"
codex_runtime_patch.patch_provider(provider, provider_out)
assert '"api_mode": "codex_app_server"' in provider_out.read_text()
session = tmp_path / "codex_app_server_session.py"
session.write_text(
codex_runtime_patch.SESSION_SIGNATURE_BEFORE
+ codex_runtime_patch.SESSION_REQUEST_BEFORE,
encoding="utf-8",
)
session_out = tmp_path / "patched/codex_app_server_session.py"
codex_runtime_patch.patch_session(session, session_out)
session_content = session_out.read_text()
assert 'turn_params["model"] = model' in session_content
assert 'turn_params["effort"] = effort' in session_content
assert '"approvalPolicy": "never"' in session_content
assert '"sandboxPolicy": {"type": "dangerFullAccess"}' in session_content
turn = tmp_path / "codex_runtime.py"
turn.write_text(
codex_runtime_patch.FALLBACK_CONTEXT_BEFORE
+ codex_runtime_patch.TURN_BEFORE,
encoding="utf-8",
)
turn_out = tmp_path / "patched/codex_runtime.py"
codex_runtime_patch.patch_turn(turn, turn_out)
turn_content = turn_out.read_text()
assert "model=str(getattr(agent" in turn_content
assert "build_cross_provider_codex_prompt" in turn_content
fallback = tmp_path / "chat_completion_helpers.py"
fallback.write_text(
codex_runtime_patch.FALLBACK_RESOLUTION_BEFORE,
encoding="utf-8",
)
fallback_out = tmp_path / "patched/chat_completion_helpers.py"
codex_runtime_patch.patch_fallback(fallback, fallback_out)
fallback_content = fallback_out.read_text()
assert 'agent.api_mode = "codex_app_server"' in fallback_content
assert "agent._codex_cross_provider_fallback = True" in fallback_content
loop = tmp_path / "conversation_loop.py"
loop.write_text(
codex_runtime_patch.FALLBACK_DISPATCH_BEFORE
+ codex_runtime_patch.RETRY_FALLBACK_DISPATCH_BEFORE
+ codex_runtime_patch.STREAM_RECOVERY_BEFORE,
encoding="utf-8",
)
loop_out = tmp_path / "patched/conversation_loop.py"
codex_runtime_patch.patch_loop(loop, loop_out)
loop_content = loop_out.read_text()
assert loop_content.count('if agent.api_mode == "codex_app_server"') == 2
assert "build_cross_provider_codex_prompt" in loop_content
retry_dispatch = loop_content.index(
"Fallback activation happens inside this retry loop"
)
api_kwargs = loop_content.find("agent._build_api_kwargs", retry_dispatch)
assert api_kwargs == -1 or retry_dispatch < api_kwargs
assert "Provider stream ended before a complete response" in loop_content
assert "_is_transport_stub" in loop_content
assert "rerouting ({truncated_tool_call_retries}/4)" in loop_content
auxiliary = tmp_path / "auxiliary_client.py"
auxiliary.write_text(
codex_runtime_patch.AUXILIARY_TOKEN_BEFORE,
encoding="utf-8",
)
auxiliary_out = tmp_path / "patched/auxiliary_client.py"
codex_runtime_patch.patch_auxiliary(auxiliary, auxiliary_out)
auxiliary_content = auxiliary_out.read_text()
assert 'os.environ.get("CODEX_HOME"' in auxiliary_content
assert 'Path(codex_home).expanduser() / "auth.json"' in auxiliary_content
assert "never creates a metered API-key lane" in auxiliary_content
def test_agent_mounts_codex_auxiliary_runtime_patch():
deployment = _agent_deployment()
pod = deployment["spec"]["template"]["spec"]
patch_init = next(
item for item in pod["initContainers"]
if item["name"] == "patch-codex-runtime"
)
assert patch_init["command"][-2:] == [
"/opt/hermes/agent/auxiliary_client.py",
"/patched/auxiliary_client.py",
]
expected_mount = {
"name": "codex-runtime-patch",
"mountPath": "/opt/hermes/agent/auxiliary_client.py",
"subPath": "auxiliary_client.py",
}
containers = {item["name"]: item for item in pod["containers"]}
for name in ("hermes", "terminal"):
assert expected_mount in containers[name]["volumeMounts"]
def test_codex_auxiliary_patch_reads_cli_token_without_copying_it(
tmp_path: Path,
monkeypatch,
):
codex_home = tmp_path / ".codex"
codex_home.mkdir()
(codex_home / "auth.json").write_text(
json.dumps({"tokens": {"access_token": "cli-access-token"}}),
encoding="utf-8",
)
monkeypatch.setenv("CODEX_HOME", str(codex_home))
source = tmp_path / "auxiliary_client.py"
source.write_text(
"import json, logging, os, time\n"
"from pathlib import Path\n"
"logger = logging.getLogger(__name__)\n"
"def read_token():\n"
" try:\n"
" raise RuntimeError('Hermes provider store intentionally empty')\n"
+ codex_runtime_patch.AUXILIARY_TOKEN_BEFORE,
encoding="utf-8",
)
destination = tmp_path / "patched/auxiliary_client.py"
codex_runtime_patch.patch_auxiliary(source, destination)
namespace: dict = {}
exec(compile(destination.read_text(), str(destination), "exec"), namespace)
assert namespace["read_token"]() == "cli-access-token"
def test_codex_runtime_patch_fails_closed_on_upstream_drift(tmp_path: Path):
source = tmp_path / "runtime_provider.py"
source.write_text("def changed():\n pass\n", encoding="utf-8")
with pytest.raises(RuntimeError, match="context changed"):
codex_runtime_patch.patch_provider(source, tmp_path / "patched.py")
def test_codex_runtime_migration_uses_owner_unsafe_mode(tmp_path: Path):
config = tmp_path / "config.yaml"
config.write_text("model: {}\n", encoding="utf-8")
codex_home = tmp_path / ".codex"
calls = []
class Report:
errors = []
@staticmethod
def summary():
return "configured"
def migrate(value, **kwargs):
calls.append((value, kwargs))
return Report()
client_config.configure_codex_runtime(config, migrate, codex_home)
assert calls[0][1]["default_permission_profile"] is None
assert calls[0][1]["codex_home"] == codex_home
content = (codex_home / "config.toml").read_text(encoding="utf-8")
assert 'approval_policy = "never"' in content
assert 'sandbox_mode = "danger-full-access"' in content
assert "default_permissions" not in content
def test_codex_owner_permissions_replace_stale_profile(tmp_path: Path):
config = tmp_path / "config.toml"
config.write_text(
'default_permissions = ":danger-no-sandbox"\n\n[features]\nhooks = true\n',
encoding="utf-8",
)
client_config.configure_codex_owner_permissions(config)
client_config.configure_codex_owner_permissions(config)
content = config.read_text(encoding="utf-8")
assert content.count(client_config.OWNER_PERMISSIONS_BEGIN) == 1
assert content.count('approval_policy = "never"') == 1
assert content.count('sandbox_mode = "danger-full-access"') == 1
assert "default_permissions" not in content
assert "[features]\nhooks = true" in content
def test_tui_gateway_patch_extends_and_bounds_agent_startup(tmp_path: Path):
source = tmp_path / "server.py"
destination = tmp_path / "patched/server.py"
source.write_text(
"import os\n\n" + tui_gateway_patch.BEFORE + "\ndef unchanged():\n pass\n",
encoding="utf-8",
)
tui_gateway_patch.patch(source, destination)
content = destination.read_text(encoding="utf-8")
assert "HERMES_TUI_AGENT_INIT_TIMEOUT_S" in content
assert 'configured = 180.0' in content
assert "return max(30.0, min(configured, 900.0))" in content
assert "timeout: float | None = None" in content
assert "ready.wait(timeout=wait_timeout)" in content
assert "def unchanged():" in content
def test_tui_gateway_patch_fails_closed_on_upstream_drift(tmp_path: Path):
source = tmp_path / "server.py"
source.write_text("def changed():\n pass\n", encoding="utf-8")
with pytest.raises(RuntimeError, match="context changed"):
tui_gateway_patch.patch(source, tmp_path / "patched.py")
def test_ttyd_clipboard_and_reconnect_patch_remain_enabled():
source = '<html><body><script>document.execCommand("copy")</script></body></html>'
content = ttyd_patch.patch_html(source)
assert 'id="atlas-ttyd-clipboard"' in content
assert "navigator.clipboard.writeText(text)" in content
assert "class AtlasRecoveringWebSocket" in content
assert "window.location.reload()" in content
assert "event.stopImmediatePropagation()" in content
def test_ttyd_patch_fails_closed_on_upstream_drift():
with pytest.raises(RuntimeError, match="context changed"):
ttyd_patch.patch_html("<html><body>changed</body></html>")