mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-06-27 11:22:03 +00:00
fix(config): surface invalid platform_toolsets instead of silently dropping tools (#38798)
A config migration (or hand-edit) that leaves an invalid toolset name in `platform_toolsets` — e.g. the #38798 corruption that rewrote `hermes-cli` to the non-existent `hermes` — silently disabled all affected tools: resolve_toolset() returns [] for an unknown name, so the agent quietly lost its tools with no error, warning, or log entry and degraded to text-only replies. Surface it loudly at two points: - After migration (migrate_config): validate platform_toolsets and record/print a warning per unknown name, with a `hermes-<platform>` suggestion when that would have been valid (the exact #38798 shape). - At runtime (_get_platform_tools): if a platform was explicitly configured but every toolset name is invalid, log a warning when tools are resolved for a session — so an ALREADY-corrupted config is caught at startup, not only on the next `hermes update`. Logic lives in a new pure, side-effect-free helper (toolset_validation.py) with validate_toolset injected, so it is unit-testable without the tool registry. Note: the original v25→v26 migration that caused the corruption no longer exists (config format is now v30; no migration step rewrites toolset names). This change is the durable defense against the silent-failure mode regardless of cause, matching the issue's "Expected: log a warning". Salvaged from #39207 by @lEWFkRAD (authorship preserved via cherry-pick). Tests: 9 helper cases (incl. the #38798 corruption shape, mixed valid/invalid, zero-tools state, non-dict/scalar/non-string) + a runtime caplog test — both the helper warning and the runtime guard mutation-verified to fail without the fix. Closes #38798. Supersedes #39581 (prevent-in-v25→v26 — that path is gone), #41006 / #40208 (repair-migration for already-corrupted configs).
This commit is contained in:
parent
0f81b0d458
commit
41ede84b93
5 changed files with 272 additions and 1 deletions
|
|
@ -5308,9 +5308,29 @@ def migrate_config(interactive: bool = True, quiet: bool = False) -> Dict[str, A
|
|||
config["mcp_servers"] = raw_mcp_servers
|
||||
save_config(config)
|
||||
|
||||
# ── Always: validate platform_toolsets after migration ──
|
||||
# A migration (or hand-edit) that leaves an invalid toolset name in
|
||||
# platform_toolsets silently disables the affected tools — resolve_toolset()
|
||||
# returns [] for an unknown name, so the agent quietly loses tools with no
|
||||
# error or warning. Surface it loudly instead. See #38798.
|
||||
try:
|
||||
from toolsets import validate_toolset
|
||||
from hermes_cli.toolset_validation import validate_platform_toolsets
|
||||
|
||||
ts_warnings = validate_platform_toolsets(
|
||||
read_raw_config().get("platform_toolsets"), validate_toolset
|
||||
)
|
||||
for w in ts_warnings:
|
||||
results["warnings"].append(w)
|
||||
if not quiet:
|
||||
print(f" ⚠ {w}")
|
||||
except Exception as _ts_val_err:
|
||||
# best-effort; never block migration on validation
|
||||
logger.debug("platform_toolsets validation skipped: %s", _ts_val_err)
|
||||
|
||||
if current_ver < latest_ver and not quiet:
|
||||
print(f"Config version: {current_ver} → {latest_ver}")
|
||||
|
||||
|
||||
# Check for missing required env vars
|
||||
missing_env = get_missing_env_vars(required_only=True)
|
||||
|
||||
|
|
|
|||
|
|
@ -34,6 +34,11 @@ from utils import base_url_hostname, is_truthy_value
|
|||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
# Platforms already warned about an all-invalid platform_toolsets list, so the
|
||||
# runtime check in _get_platform_tools warns once per platform instead of on
|
||||
# every tool resolution for a persistently-corrupt config (#38798).
|
||||
_warned_invalid_platform_toolsets: Set[str] = set()
|
||||
|
||||
PROJECT_ROOT = Path(__file__).parent.parent.resolve()
|
||||
|
||||
|
||||
|
|
@ -1611,6 +1616,31 @@ def _get_platform_tools(
|
|||
disabled_set = {str(ts) for ts in disabled_toolsets}
|
||||
enabled_toolsets -= disabled_set
|
||||
|
||||
# #38798: if this platform was explicitly configured but every toolset name
|
||||
# is invalid (e.g. a migration or hand-edit left `hermes` instead of
|
||||
# `hermes-cli`), resolve_toolset() returns [] for each and the platform ends
|
||||
# up with no native tools — silently, with no error. Surface it at the point
|
||||
# tools are resolved for a session so an already-corrupted config is caught
|
||||
# at runtime, not only during the next `hermes update`/`hermes doctor`.
|
||||
_explicit = platform_toolsets.get(platform)
|
||||
if isinstance(_explicit, list) and _explicit:
|
||||
from toolsets import validate_toolset
|
||||
|
||||
_named = [str(t) for t in _explicit if isinstance(t, str) and t]
|
||||
if (
|
||||
_named
|
||||
and not any(validate_toolset(t) for t in _named)
|
||||
and platform not in _warned_invalid_platform_toolsets
|
||||
):
|
||||
_warned_invalid_platform_toolsets.add(platform)
|
||||
logger.warning(
|
||||
"platform '%s' has no valid toolsets configured (unknown "
|
||||
"name(s): %s) - tools will be unavailable. Run `hermes tools` "
|
||||
"to reconfigure. See issue #38798.",
|
||||
platform,
|
||||
", ".join(_named),
|
||||
)
|
||||
|
||||
return enabled_toolsets
|
||||
|
||||
|
||||
|
|
|
|||
74
hermes_cli/toolset_validation.py
Normal file
74
hermes_cli/toolset_validation.py
Normal file
|
|
@ -0,0 +1,74 @@
|
|||
"""Validation for the ``platform_toolsets`` config section.
|
||||
|
||||
Pure, side-effect-free helpers so the logic is unit-testable without importing
|
||||
the tool registry or launching Hermes (mirrors the decoupled-helper pattern used
|
||||
elsewhere in the CLI).
|
||||
|
||||
Motivated by #38798: a config migration silently rewrote the valid toolset name
|
||||
``hermes-cli`` to the non-existent ``hermes``. ``resolve_toolset('hermes')``
|
||||
returns an empty list, so every tool silently disappeared with no error, warning,
|
||||
or log entry — the agent degraded to text-only replies and the cause took
|
||||
significant debugging to find. Surfacing invalid toolset names (and the
|
||||
zero-tools end state) loudly turns that silent failure into an actionable one.
|
||||
"""
|
||||
|
||||
from typing import Callable, Dict, List
|
||||
|
||||
|
||||
def validate_platform_toolsets(
|
||||
platform_toolsets: object,
|
||||
is_valid_toolset: Callable[[str], bool],
|
||||
) -> List[str]:
|
||||
"""Return human-readable warnings for a ``platform_toolsets`` mapping.
|
||||
|
||||
Two failure modes are reported:
|
||||
|
||||
1. A toolset name that ``is_valid_toolset`` rejects — usually a corrupted or
|
||||
renamed entry. When ``hermes-<platform>`` would have been valid (the exact
|
||||
#38798 shape, where ``cli`` held ``hermes`` instead of ``hermes-cli``),
|
||||
the warning includes that as a suggestion.
|
||||
2. The mapping is non-empty but resolves to *zero* valid toolsets, so the
|
||||
agent would start with no tools at all.
|
||||
|
||||
``is_valid_toolset`` is injected (normally :func:`toolsets.validate_toolset`)
|
||||
so this function performs no imports or I/O and is testable in isolation.
|
||||
|
||||
Args:
|
||||
platform_toolsets: The raw ``platform_toolsets`` value from config. Only
|
||||
``dict`` values carry toolset entries; anything else yields no
|
||||
warnings (nothing to validate).
|
||||
is_valid_toolset: Predicate returning ``True`` for a known toolset name.
|
||||
|
||||
Returns:
|
||||
A list of warning strings (empty when everything is valid).
|
||||
"""
|
||||
warnings: List[str] = []
|
||||
if not isinstance(platform_toolsets, dict) or not platform_toolsets:
|
||||
return warnings
|
||||
|
||||
valid_count = 0
|
||||
for platform, raw in platform_toolsets.items():
|
||||
names = raw if isinstance(raw, list) else [raw]
|
||||
for name in names:
|
||||
if not isinstance(name, str) or not name:
|
||||
continue
|
||||
if is_valid_toolset(name):
|
||||
valid_count += 1
|
||||
continue
|
||||
suggestion = f"hermes-{platform}"
|
||||
hint = (
|
||||
f" — did you mean '{suggestion}'?"
|
||||
if is_valid_toolset(suggestion)
|
||||
else ""
|
||||
)
|
||||
warnings.append(
|
||||
f"platform '{platform}' references unknown toolset "
|
||||
f"'{name}'{hint}"
|
||||
)
|
||||
|
||||
if valid_count == 0:
|
||||
warnings.append(
|
||||
"platform_toolsets resolves to zero valid toolsets — the agent will "
|
||||
"have no tools. Run `hermes tools` to reconfigure."
|
||||
)
|
||||
return warnings
|
||||
|
|
@ -1,5 +1,6 @@
|
|||
"""Tests for hermes_cli.tools_config platform tool persistence."""
|
||||
|
||||
import logging
|
||||
from types import SimpleNamespace
|
||||
from unittest.mock import patch
|
||||
|
||||
|
|
@ -58,6 +59,61 @@ def test_agent_disabled_toolsets_with_explicit_platform_config():
|
|||
assert "terminal" in enabled
|
||||
|
||||
|
||||
def test_all_invalid_platform_toolsets_logs_runtime_warning(caplog):
|
||||
"""#38798: an explicit platform config whose toolset names are all invalid
|
||||
(e.g. 'hermes' instead of 'hermes-cli') must warn at resolve time so an
|
||||
already-corrupted config is caught at runtime, not just during migration."""
|
||||
import hermes_cli.tools_config as _tc
|
||||
# The runtime warning fires once per platform per process; clear the guard
|
||||
# so this test is deterministic regardless of prior resolutions.
|
||||
_tc._warned_invalid_platform_toolsets.discard("cli")
|
||||
config = {"platform_toolsets": {"cli": ["hermes"]}}
|
||||
|
||||
with caplog.at_level(logging.WARNING, logger="hermes_cli.tools_config"):
|
||||
_get_platform_tools(config, "cli")
|
||||
|
||||
warnings = [r.getMessage() for r in caplog.records if r.levelno >= logging.WARNING]
|
||||
assert any("#38798" in m and "hermes" in m for m in warnings), warnings
|
||||
|
||||
|
||||
def test_invalid_platform_toolsets_runtime_warning_fires_once(caplog):
|
||||
"""The runtime warning is deduped per platform — a persistently-corrupt
|
||||
config must not spam an identical warning on every tool resolution."""
|
||||
import hermes_cli.tools_config as _tc
|
||||
_tc._warned_invalid_platform_toolsets.discard("cli")
|
||||
config = {"platform_toolsets": {"cli": ["hermes"]}}
|
||||
|
||||
with caplog.at_level(logging.WARNING, logger="hermes_cli.tools_config"):
|
||||
_get_platform_tools(config, "cli")
|
||||
_get_platform_tools(config, "cli")
|
||||
_get_platform_tools(config, "cli")
|
||||
|
||||
hits = [r for r in caplog.records if "#38798" in r.getMessage()]
|
||||
assert len(hits) == 1, f"expected exactly one warning, got {len(hits)}"
|
||||
|
||||
|
||||
def test_valid_platform_toolsets_no_runtime_warning(caplog):
|
||||
"""A correctly-configured platform must not emit the #38798 warning."""
|
||||
config = {"platform_toolsets": {"cli": ["hermes-cli"]}}
|
||||
|
||||
with caplog.at_level(logging.WARNING, logger="hermes_cli.tools_config"):
|
||||
_get_platform_tools(config, "cli")
|
||||
|
||||
assert not any("#38798" in r.getMessage() for r in caplog.records)
|
||||
|
||||
|
||||
def test_partially_valid_platform_toolsets_no_runtime_warning(caplog):
|
||||
"""When at least one configured toolset is valid, tools still resolve, so
|
||||
the runtime zero-tools warning must not fire (the migration-time check still
|
||||
flags the individual bad name)."""
|
||||
config = {"platform_toolsets": {"cli": ["hermes-cli", "bogus"]}}
|
||||
|
||||
with caplog.at_level(logging.WARNING, logger="hermes_cli.tools_config"):
|
||||
_get_platform_tools(config, "cli")
|
||||
|
||||
assert not any("#38798" in r.getMessage() for r in caplog.records)
|
||||
|
||||
|
||||
def test_agent_disabled_toolsets_empty_list_is_noop():
|
||||
"""Empty or missing disabled_toolsets should not change behavior."""
|
||||
config_empty = {"agent": {"disabled_toolsets": []}}
|
||||
|
|
|
|||
91
tests/hermes_cli/test_toolset_validation.py
Normal file
91
tests/hermes_cli/test_toolset_validation.py
Normal file
|
|
@ -0,0 +1,91 @@
|
|||
"""Unit tests for hermes_cli.toolset_validation (see #38798).
|
||||
|
||||
Pure logic — the validity predicate is injected, so these tests need neither the
|
||||
tool registry nor a running Hermes.
|
||||
"""
|
||||
|
||||
import pytest
|
||||
|
||||
from hermes_cli.toolset_validation import validate_platform_toolsets
|
||||
|
||||
# A representative set of real toolset names. `hermes` is deliberately absent —
|
||||
# that is the corruption #38798 reported (`hermes-cli` rewritten to `hermes`).
|
||||
_KNOWN = {
|
||||
"hermes-cli",
|
||||
"hermes-telegram",
|
||||
"hermes-discord",
|
||||
"terminal",
|
||||
"web",
|
||||
}
|
||||
|
||||
|
||||
def _is_valid(name):
|
||||
return name in _KNOWN
|
||||
|
||||
|
||||
def test_valid_config_produces_no_warnings():
|
||||
cfg = {"cli": ["hermes-cli"], "telegram": ["hermes-telegram"]}
|
||||
assert validate_platform_toolsets(cfg, _is_valid) == []
|
||||
|
||||
|
||||
def test_38798_corruption_warns_and_suggests_correct_name():
|
||||
# The exact reported shape: cli holds 'hermes' instead of 'hermes-cli'.
|
||||
warnings = validate_platform_toolsets({"cli": ["hermes"]}, _is_valid)
|
||||
unknown = [w for w in warnings if "unknown toolset 'hermes'" in w]
|
||||
assert len(unknown) == 1
|
||||
# Actionable: points at the valid name the entry should have been.
|
||||
assert "did you mean 'hermes-cli'?" in unknown[0]
|
||||
# And the zero-valid-toolsets safety net fires.
|
||||
assert any("zero valid toolsets" in w for w in warnings)
|
||||
|
||||
|
||||
def test_mixed_valid_and_invalid_flags_only_the_invalid():
|
||||
cfg = {"cli": ["hermes-cli"], "discord": ["bogus"]}
|
||||
warnings = validate_platform_toolsets(cfg, _is_valid)
|
||||
# One valid entry exists, so no zero-valid warning.
|
||||
assert not any("zero valid toolsets" in w for w in warnings)
|
||||
assert len(warnings) == 1
|
||||
assert "platform 'discord'" in warnings[0]
|
||||
assert "unknown toolset 'bogus'" in warnings[0]
|
||||
|
||||
|
||||
def test_unknown_without_valid_platform_default_omits_suggestion():
|
||||
# hermes-mystery is not a known toolset, so no "did you mean" hint.
|
||||
warnings = validate_platform_toolsets({"mystery": ["nope"]}, _is_valid)
|
||||
unknown = [w for w in warnings if "unknown toolset 'nope'" in w]
|
||||
assert len(unknown) == 1
|
||||
assert "did you mean" not in unknown[0]
|
||||
|
||||
|
||||
@pytest.mark.parametrize("value", [None, {}, [], "hermes-cli", 42])
|
||||
def test_non_dict_or_empty_yields_no_warnings(value):
|
||||
assert validate_platform_toolsets(value, _is_valid) == []
|
||||
|
||||
|
||||
def test_scalar_toolset_value_is_accepted():
|
||||
# Some configs store the toolset as a bare string rather than a list.
|
||||
assert validate_platform_toolsets({"cli": "hermes-cli"}, _is_valid) == []
|
||||
|
||||
|
||||
def test_non_string_entries_are_skipped_not_counted_invalid():
|
||||
cfg = {"cli": [None, 123, "hermes-cli"]}
|
||||
# The junk entries are ignored; the valid one keeps it from being "zero".
|
||||
assert validate_platform_toolsets(cfg, _is_valid) == []
|
||||
|
||||
|
||||
def test_all_invalid_reports_each_and_the_zero_state():
|
||||
cfg = {"cli": ["hermes"], "discord": ["hermes"]}
|
||||
warnings = validate_platform_toolsets(cfg, _is_valid)
|
||||
assert sum("unknown toolset" in w for w in warnings) == 2
|
||||
assert any("zero valid toolsets" in w for w in warnings)
|
||||
|
||||
|
||||
def test_real_validate_toolset_treats_hermes_cli_valid_and_hermes_invalid():
|
||||
# Ties the helper to reality: the canonical registry check agrees that
|
||||
# `hermes-cli` is the real toolset and `hermes` is not (the #38798 crux).
|
||||
from toolsets import validate_toolset
|
||||
|
||||
assert validate_toolset("hermes-cli") is True
|
||||
assert validate_toolset("hermes") is False
|
||||
warnings = validate_platform_toolsets({"cli": ["hermes"]}, validate_toolset)
|
||||
assert any("did you mean 'hermes-cli'?" in w for w in warnings)
|
||||
Loading…
Add table
Add a link
Reference in a new issue