refactor: move standalone scripts to scripts/ directory

Move batch_runner, trajectory_compressor, mini_swe_runner, and rl_cli
from the project root into scripts/, update all imports, logger names,
pyproject.toml, and downstream test references.

cli.py

@@ -7000,7 +7000,7 @@ class HermesCLI:
                 logging.getLogger(noisy).setLevel(logging.WARNING)
         else:
             logging.getLogger().setLevel(logging.INFO)
-            for quiet_logger in ('tools', 'run_agent', 'trajectory_compressor', 'cron', 'hermes_cli'):
+            for quiet_logger in ('tools', 'run_agent', 'scripts.trajectory_compressor', 'cron', 'hermes_cli'):
                 logging.getLogger(quiet_logger).setLevel(logging.ERROR)
 
     def _show_insights(self, command: str = "/insights"):

datagen-config-examples/run_browser_tasks.sh

@@ -29,7 +29,7 @@ echo "📝 Logging to: $LOG_FILE"
 # Point to the example dataset in this directory
 SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
 
-python batch_runner.py \
+python scripts/batch_runner.py \
   --dataset_file="$SCRIPT_DIR/example_browser_tasks.jsonl" \
   --batch_size=5 \
   --run_name="browser_tasks_example" \

datagen-config-examples/web_research.yaml

@@ -4,7 +4,7 @@
 # Generates tool-calling trajectories for multi-step web research tasks.
 #
 # Usage:
-#   python batch_runner.py \
+#   python scripts/batch_runner.py \
 #     --config datagen-config-examples/web_research.yaml \
 #     --run_name web_research_v1
 

hermes_logging.py

@@ -142,7 +142,7 @@ class _ComponentFilter(logging.Filter):
 # Used by _ComponentFilter and exposed for ``hermes logs --component``.
 COMPONENT_PREFIXES = {
     "gateway": ("gateway",),
-    "agent": ("agent", "run_agent", "model_tools", "batch_runner"),
+    "agent": ("agent", "run_agent", "model_tools", "scripts.batch_runner"),
     "tools": ("tools",),
     "cli": ("hermes_cli", "cli"),
     "cron": ("cron",),

pyproject.toml

@@ -120,13 +120,13 @@ hermes-agent = "run_agent:main"
 hermes-acp = "acp_adapter.entry:main"
 
 [tool.setuptools]
-py-modules = ["run_agent", "model_tools", "toolsets", "batch_runner", "trajectory_compressor", "toolset_distributions", "cli", "hermes_constants", "hermes_state", "hermes_time", "hermes_logging", "rl_cli", "utils"]
+py-modules = ["run_agent", "model_tools", "toolsets", "toolset_distributions", "cli", "hermes_constants", "hermes_state", "hermes_time", "hermes_logging", "utils"]
 
 [tool.setuptools.package-data]
 hermes_cli = ["web_dist/**/*"]
 
 [tool.setuptools.packages.find]
-include = ["agent", "tools", "tools.*", "hermes_cli", "gateway", "gateway.*", "tui_gateway", "tui_gateway.*", "cron", "acp_adapter", "plugins", "plugins.*"]
+include = ["agent", "tools", "tools.*", "hermes_cli", "gateway", "gateway.*", "tui_gateway", "tui_gateway.*", "cron", "acp_adapter", "plugins", "plugins.*", "scripts"]
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]

run_agent.py

@@ -1041,7 +1041,7 @@ class AIAgent:
                 for quiet_logger in [
                     'tools',               # all tools.* (terminal, browser, web, file, etc.)
                     'run_agent',            # agent runner internals
-                    'trajectory_compressor',
+                    'scripts.trajectory_compressor',
                     'cron',                 # scheduler (only relevant in daemon mode)
                     'hermes_cli',           # CLI helpers
                 ]:

scripts/batch_runner.py

@@ -20,9 +20,13 @@ Usage:
     python batch_runner.py --dataset_file=data.jsonl --batch_size=10 --run_name=my_run --distribution=image_gen
 """
 
+import os
+import sys
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
 import json
 import logging
-import os
 import time
 from pathlib import Path
 from typing import List, Dict, Any, Optional, Tuple

scripts/mini_swe_runner.py

@@ -26,10 +26,13 @@ Usage:
     python mini_swe_runner.py --prompts_file prompts.jsonl --output_file trajectories.jsonl --env docker
 """
 
-import json
-import logging
 import os
 import sys
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import json
+import logging
 import time
 import uuid
 from datetime import datetime

scripts/rl_cli.py

@@ -19,9 +19,12 @@ Environment Variables:
     OPENROUTER_API_KEY: API key for OpenRouter (required for agent)
 """
 
-import asyncio
 import os
 import sys
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import asyncio
 from pathlib import Path
 
 import fire
@@ -32,7 +35,7 @@ from hermes_constants import get_hermes_home, OPENROUTER_BASE_URL
 # Load .env from ~/.hermes/.env first, then project root as dev fallback.
 # User-managed env files should override stale shell exports on restart.
 _hermes_home = get_hermes_home()
-_project_env = Path(__file__).parent / '.env'
+_project_env = Path(__file__).parent.parent / '.env'
 
 from hermes_cli.env_loader import load_hermes_dotenv
 

scripts/sample_and_compress.py

@@ -267,7 +267,7 @@ def run_compression(input_dir: Path, output_dir: Path, config_path: str):
     # Import the compressor
     import sys
     sys.path.insert(0, str(Path(__file__).parent.parent))
-    from trajectory_compressor import TrajectoryCompressor, CompressionConfig
+    from scripts.trajectory_compressor import TrajectoryCompressor, CompressionConfig
     
     print(f"\n🗜️  Running trajectory compression...")
     print(f"   Input: {input_dir}")

scripts/trajectory_compressor.py

@@ -30,8 +30,12 @@ Usage:
     python trajectory_compressor.py --input=data/my_run --sample_percent=10
 """
 
-import json
 import os
+import sys
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+import json
 import time
 import yaml
 import logging
@@ -52,7 +56,7 @@ from agent.retry_utils import jittered_backoff
 from hermes_cli.env_loader import load_hermes_dotenv
 
 _hermes_home = get_hermes_home()
-_project_env = Path(__file__).parent / ".env"
+_project_env = Path(__file__).parent.parent / ".env"
 load_hermes_dotenv(hermes_home=_hermes_home, project_env=_project_env)
 
 

tests/hermes_cli/test_arcee_provider.py

@@ -164,7 +164,7 @@ class TestArceeURLMapping:
         assert "arceeai" in _PROVIDER_PREFIXES
 
     def test_trajectory_compressor_detects_arcee(self):
-        import trajectory_compressor as tc
+        import scripts.trajectory_compressor as tc
         comp = tc.TrajectoryCompressor.__new__(tc.TrajectoryCompressor)
         comp.config = types.SimpleNamespace(base_url="https://api.arcee.ai/api/v1")
         assert comp._detect_provider() == "arcee"

tests/integration/test_batch_runner.py

@@ -104,7 +104,7 @@ def main():
     test_file = create_test_dataset()
     
     print(f"\n📝 To run the test manually:")
-    print(f"   python batch_runner.py \\")
+    print(f"   python scripts/batch_runner.py \\")
     print(f"       --dataset_file={test_file} \\")
     print(f"       --batch_size=2 \\")
     print(f"       --run_name={run_name} \\")
@@ -112,7 +112,7 @@ def main():
     print(f"       --num_workers=2")
     
     print(f"\n💡 Or test with different distributions:")
-    print(f"   python batch_runner.py --list_distributions")
+    print(f"   python scripts/batch_runner.py --list_distributions")
     
     print(f"\n🔍 After running, you can verify output with:")
     print(f"   python tests/test_batch_runner.py --verify")

tests/integration/test_checkpoint_resumption.py

@@ -30,7 +30,7 @@ from pathlib import Path
 from typing import List, Dict, Any
 import traceback
 
-# Add project root to path to import batch_runner
+# Add project root to path to import scripts.batch_runner
 sys.path.insert(0, str(Path(__file__).parent.parent.parent))
 
 
@@ -135,7 +135,7 @@ def test_current_implementation():
         shutil.rmtree(output_dir)
     
     # Import here to avoid issues if module changes
-    from batch_runner import BatchRunner
+    from scripts.batch_runner import BatchRunner
     
     checkpoint_file = output_dir / "checkpoint.json"
     
@@ -229,7 +229,7 @@ def test_interruption_and_resume():
     if output_dir.exists():
         shutil.rmtree(output_dir)
     
-    from batch_runner import BatchRunner
+    from scripts.batch_runner import BatchRunner
     
     checkpoint_file = output_dir / "checkpoint.json"
     

tests/test_batch_runner_checkpoint.py

@@ -8,11 +8,7 @@ from unittest.mock import patch, MagicMock
 
 import pytest
 
-# batch_runner uses relative imports, ensure project root is on path
+from scripts.batch_runner import BatchRunner, _process_batch_worker
-import sys
-sys.path.insert(0, str(Path(__file__).parent.parent))
-
-from batch_runner import BatchRunner, _process_batch_worker
 
 
 @pytest.fixture
@@ -173,7 +169,7 @@ class TestBatchWorkerResumeBehavior:
             "toolsets_used": [],
         }
 
-        monkeypatch.setattr("batch_runner._process_single_prompt", lambda *args, **kwargs: prompt_result)
+        monkeypatch.setattr("scripts.batch_runner._process_single_prompt", lambda *args, **kwargs: prompt_result)
 
         result = _process_batch_worker((
             1,

tests/test_mini_swe_runner.py

@@ -14,7 +14,7 @@ def test_run_task_kimi_omits_temperature():
         )
         mock_openai.return_value = client
 
-        from mini_swe_runner import MiniSWERunner
+        from scripts.mini_swe_runner import MiniSWERunner
 
         runner = MiniSWERunner(
             model="kimi-for-coding",
@@ -42,7 +42,7 @@ def test_run_task_public_moonshot_kimi_k2_5_omits_temperature():
         )
         mock_openai.return_value = client
 
-        from mini_swe_runner import MiniSWERunner
+        from scripts.mini_swe_runner import MiniSWERunner
 
         runner = MiniSWERunner(
             model="kimi-k2.5",

tests/test_trajectory_compressor.py

@@ -9,7 +9,7 @@ from unittest.mock import AsyncMock, patch, MagicMock
 
 import pytest
 
-from trajectory_compressor import (
+from scripts.trajectory_compressor import (
     CompressionConfig,
     TrajectoryMetrics,
     AggregateMetrics,
@@ -25,8 +25,8 @@ def test_import_loads_env_from_hermes_home(tmp_path, monkeypatch):
     monkeypatch.setenv("HERMES_HOME", str(home))
     monkeypatch.delenv("OPENROUTER_API_KEY", raising=False)
 
-    sys.modules.pop("trajectory_compressor", None)
+    sys.modules.pop("scripts.trajectory_compressor", None)
-    importlib.import_module("trajectory_compressor")
+    importlib.import_module("scripts.trajectory_compressor")
 
     assert os.getenv("OPENROUTER_API_KEY") == "from-hermes-home"
 

tests/test_trajectory_compressor_async.py

@@ -22,7 +22,7 @@ class TestAsyncClientLazyCreation:
 
     def test_async_client_none_after_init(self):
         """async_client should be None after __init__ (not eagerly created)."""
-        from trajectory_compressor import TrajectoryCompressor
+        from scripts.trajectory_compressor import TrajectoryCompressor
 
         comp = TrajectoryCompressor.__new__(TrajectoryCompressor)
         comp.config = MagicMock()
@@ -36,7 +36,7 @@ class TestAsyncClientLazyCreation:
 
     def test_get_async_client_creates_new_client(self):
         """_get_async_client() should create a fresh AsyncOpenAI instance."""
-        from trajectory_compressor import TrajectoryCompressor
+        from scripts.trajectory_compressor import TrajectoryCompressor
 
         comp = TrajectoryCompressor.__new__(TrajectoryCompressor)
         comp.config = MagicMock()
@@ -57,7 +57,7 @@ class TestAsyncClientLazyCreation:
     def test_get_async_client_creates_fresh_each_call(self):
         """Each call to _get_async_client() creates a NEW client instance,
         so it binds to the current event loop."""
-        from trajectory_compressor import TrajectoryCompressor
+        from scripts.trajectory_compressor import TrajectoryCompressor
 
         comp = TrajectoryCompressor.__new__(TrajectoryCompressor)
         comp.config = MagicMock()
@@ -91,7 +91,7 @@ class TestSourceLineVerification:
     def _read_file() -> str:
         import os
         base = os.path.dirname(os.path.dirname(__file__))
-        with open(os.path.join(base, "trajectory_compressor.py")) as f:
+        with open(os.path.join(base, "scripts", "trajectory_compressor.py")) as f:
             return f.read()
 
     def test_no_eager_async_openai_in_init(self):
@@ -119,7 +119,7 @@ class TestSourceLineVerification:
 @pytest.mark.asyncio
 async def test_generate_summary_async_kimi_omits_temperature():
     """Kimi models should have temperature omitted — server manages it."""
-    from trajectory_compressor import CompressionConfig, TrajectoryCompressor, TrajectoryMetrics
+    from scripts.trajectory_compressor import CompressionConfig, TrajectoryCompressor, TrajectoryMetrics
 
     config = CompressionConfig(
         summarization_model="kimi-for-coding",
@@ -147,7 +147,7 @@ async def test_generate_summary_async_kimi_omits_temperature():
 @pytest.mark.asyncio
 async def test_generate_summary_async_public_moonshot_kimi_k2_5_omits_temperature():
     """kimi-k2.5 on the public Moonshot API should not get a forced temperature."""
-    from trajectory_compressor import CompressionConfig, TrajectoryCompressor, TrajectoryMetrics
+    from scripts.trajectory_compressor import CompressionConfig, TrajectoryCompressor, TrajectoryMetrics
 
     config = CompressionConfig(
         summarization_model="kimi-k2.5",
@@ -176,7 +176,7 @@ async def test_generate_summary_async_public_moonshot_kimi_k2_5_omits_temperatur
 @pytest.mark.asyncio
 async def test_generate_summary_async_public_moonshot_cn_kimi_k2_5_omits_temperature():
     """kimi-k2.5 on api.moonshot.cn should not get a forced temperature."""
-    from trajectory_compressor import CompressionConfig, TrajectoryCompressor, TrajectoryMetrics
+    from scripts.trajectory_compressor import CompressionConfig, TrajectoryCompressor, TrajectoryMetrics
 
     config = CompressionConfig(
         summarization_model="kimi-k2.5",

tests/tools/test_config_null_guard.py

@@ -87,7 +87,7 @@ class TestTrajectoryCompressorNullGuard:
 
     def test_null_base_url_does_not_crash(self):
         """base_url=None should not crash _detect_provider()."""
-        from trajectory_compressor import CompressionConfig, TrajectoryCompressor
+        from scripts.trajectory_compressor import CompressionConfig, TrajectoryCompressor
 
         config = CompressionConfig()
         config.base_url = None
@@ -101,7 +101,7 @@ class TestTrajectoryCompressorNullGuard:
 
     def test_config_loading_null_base_url_keeps_default(self):
         """YAML ``summarization: {base_url: null}`` should keep default."""
-        from trajectory_compressor import CompressionConfig
+        from scripts.trajectory_compressor import CompressionConfig
         from hermes_constants import OPENROUTER_BASE_URL
 
         config = CompressionConfig()