mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-31 19:16:29 +00:00
39975613b1
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.
82 lines
2.9 KiB
Python
82 lines
2.9 KiB
Python
"""Tests for GNU long-option abbreviation bypass in DANGEROUS_PATTERNS.
|
|
|
|
GNU tools accept unique long-option prefix abbreviations at runtime
|
|
(e.g. ``chown --recur`` resolves to ``chown --recursive``). Two approval
|
|
patterns matched only the full flag name and could be evaded by passing a
|
|
valid abbreviation:
|
|
|
|
chown --recursive → --recur[a-z]*
|
|
git push --force → --forc[a-z]*
|
|
|
|
The other long-flag patterns (rm/chmod/sed) were already covered on every
|
|
abbreviation by sibling short-flag / target patterns, so this file only
|
|
asserts the two gaps that were genuinely open plus the relevant regression
|
|
guards.
|
|
"""
|
|
|
|
import pytest
|
|
|
|
from tools.approval import detect_dangerous_command
|
|
|
|
|
|
class TestChownRecursiveLongOptionAbbreviation:
|
|
"""chown --recur* abbreviations targeting root must be caught.
|
|
|
|
On main the bare ``--recur root`` form is caught only by an accidental
|
|
case-insensitive ``r`` → ``R`` overlap with the short-flag pattern; longer
|
|
abbreviations like ``--recurs``/``--recursi`` break that overlap and
|
|
slipped through before the prefix change.
|
|
"""
|
|
|
|
def test_chown_recursive_full_still_detected(self):
|
|
dangerous, _, desc = detect_dangerous_command("chown --recursive root /etc")
|
|
assert dangerous is True
|
|
assert "chown" in desc.lower() or "root" in desc.lower()
|
|
|
|
|
|
def test_chown_recur_non_root_not_flagged(self):
|
|
"""--recur* chown to a non-root user must not be flagged."""
|
|
dangerous, _, _ = detect_dangerous_command("chown --recur nobody /opt/app")
|
|
assert dangerous is False
|
|
|
|
|
|
class TestGitPushForceLongOptionAbbreviation:
|
|
"""git push --forc* abbreviations must be caught.
|
|
|
|
The short ``-f`` pattern does not catch ``--forc`` (the ``\\b`` after the
|
|
``f`` does not match mid-word), so abbreviated long forms slipped through.
|
|
"""
|
|
|
|
def test_git_push_force_full_still_detected(self):
|
|
dangerous, _, desc = detect_dangerous_command("git push --force origin main")
|
|
assert dangerous is True
|
|
assert "force" in desc.lower()
|
|
|
|
|
|
def test_git_push_short_f_still_detected(self):
|
|
"""Existing -f pattern must not regress."""
|
|
dangerous, _, _ = detect_dangerous_command("git push -f origin main")
|
|
assert dangerous is True
|
|
|
|
|
|
def test_git_push_set_upstream_not_flagged(self):
|
|
dangerous, _, _ = detect_dangerous_command(
|
|
"git push --set-upstream origin feature"
|
|
)
|
|
assert dangerous is False
|
|
|
|
|
|
class TestFullFormRegressions:
|
|
"""The two changed long-flag patterns must still detect their full form."""
|
|
|
|
@pytest.mark.parametrize(
|
|
"cmd",
|
|
[
|
|
"chown --recursive root /etc",
|
|
"git push --force origin main",
|
|
],
|
|
)
|
|
def test_full_form_still_detected(self, cmd):
|
|
dangerous, key, _ = detect_dangerous_command(cmd)
|
|
assert dangerous is True, f"Full-form long flag not detected in: {cmd!r}"
|
|
assert key is not None
|