Skip to content

Self-improving AI Employees for everyone

fix: RLM service compatibility and Python 3.13 constraint

trohitg0ca3e7a8fa4b
diff --git a/docs-internal/rlm_service.md b/docs-internal/rlm_service.md
index 222478aa..60427648 100644
--- a/docs-internal/rlm_service.md
+++ b/docs-internal/rlm_service.md
@@ -230,8 +230,8 @@ return await ai_service.rlm_service.execute(node_id, parameters, ...)
| `server/services/node_executor.py` | `'rlm_agent': partial(handle_rlm_agent, ...)` in `_build_handler_registry()` |
| `server/services/handlers/__init__.py` | `from .rlm import handle_rlm_agent` |
| `server/services/ai.py` | `self.rlm_service = RLMService(auth=self.auth)` in `AIService.__init__()` |
-| `server/requirements.txt` | `rlms>=0.1.2` |
-| `server/pyproject.toml` | `"rlms>=0.1.2"` in `[project] dependencies` |
+| `server/requirements.txt` | `rlms>=0.1.1` |
+| `server/pyproject.toml` | `"rlms>=0.1.1"` in `[project] dependencies` |
---
diff --git a/server/pyproject.toml b/server/pyproject.toml
index 62c1d6fb..045fabf7 100644
--- a/server/pyproject.toml
+++ b/server/pyproject.toml
@@ -2,7 +2,7 @@
name = "machinaos-server"
version = "0.1.0"
description = "MachinaOS workflow automation backend"
-requires-python = ">=3.11"
+requires-python = ">=3.11,<3.13"
dependencies = [
# Core
"fastapi>=0.115.0",
@@ -86,7 +86,7 @@ dependencies = [
"temporalio>=1.7.0",
# RLM (Recursive Language Models) agent
- "rlms>=0.1.2",
+ "rlms>=0.1.1",
]
[project.optional-dependencies]
diff --git a/server/requirements.txt b/server/requirements.txt
index 7837bc60..1fde600e 100644
--- a/server/requirements.txt
+++ b/server/requirements.txt
@@ -103,5 +103,5 @@ pypdf>=4.0.0
# unstructured>=0.16.0 # Optional: Multi-format document parsing
# RLM (Recursive Language Models) agent
-rlms>=0.1.2
+rlms>=0.1.1
diff --git a/server/services/rlm/adapters.py b/server/services/rlm/adapters.py
index 3becf947..706bff92 100644
--- a/server/services/rlm/adapters.py
+++ b/server/services/rlm/adapters.py
@@ -98,14 +98,14 @@ class ToolBridgeAdapter:
}
@staticmethod
- def bridge(tool_data: Optional[List[Dict[str, Any]]], context: Optional[Dict] = None) -> Dict[str, Dict]:
+ def bridge(tool_data: Optional[List[Dict[str, Any]]], context: Optional[Dict] = None, loop: Optional[asyncio.AbstractEventLoop] = None) -> Dict[str, Dict]:
from constants import AI_AGENT_TYPES, AI_CHAT_MODEL_TYPES
from services.handlers.tools import execute_tool
if not tool_data:
return {}
- main_loop = asyncio.get_event_loop()
+ main_loop = loop or asyncio.get_event_loop()
tools = {}
for tool_info in tool_data:
diff --git a/server/services/rlm/service.py b/server/services/rlm/service.py
index 9babd0a9..ba236d6f 100644
--- a/server/services/rlm/service.py
+++ b/server/services/rlm/service.py
@@ -56,14 +56,14 @@ class RLMService:
system_message = parameters.get('systemMessage', '')
# Skill injection (reuse ai.py:833 _build_skill_system_prompt)
+ # Collected as supplementary context -- appended to RLM prompt, never replaces it
from services.ai import _build_skill_system_prompt
+ supplementary_context = ""
skill_prompt, has_personality = _build_skill_system_prompt(
skill_data, log_prefix="[RLM]"
)
if skill_prompt:
- system_message = skill_prompt if has_personality else (
- f"{system_message}\n\n{skill_prompt}" if system_message else skill_prompt
- )
+ supplementary_context += f"\n\n## Skill Instructions\n{skill_prompt}"
# Options flattening (same as execute_chat_agent line 1995)
options = parameters.get('options', {})
@@ -88,10 +88,10 @@ class RLMService:
"provider": provider, "model": model,
})
- # === Memory injection ===
+ # === Memory injection (appended to supplementary context, not system prompt) ===
if memory_data and memory_data.get('memory_content'):
memory_text = memory_data['memory_content']
- system_message = (system_message or "") + f"\n\n## Conversation History\n{memory_text}"
+ supplementary_context += f"\n\n## Conversation History\n{memory_text}"
# === Adapter: Map provider to RLM backend ===
backend, backend_kwargs = BackendAdapter.adapt(provider, model, api_key)
@@ -105,7 +105,8 @@ class RLMService:
await broadcast_status("building_tools", {
"message": "Bridging connected tools..."
})
- custom_tools = ToolBridgeAdapter.bridge(tool_data, context)
+ running_loop = asyncio.get_running_loop()
+ custom_tools = ToolBridgeAdapter.bridge(tool_data, context, loop=running_loop)
# === RLM-specific parameters ===
max_iterations = int(flattened.get('maxIterations', DEFAULT_MAX_ITERATIONS))
@@ -122,6 +123,20 @@ class RLMService:
"tool_count": len(custom_tools),
})
+ # Build augmented prompt: MachinaOS context prepended to user prompt.
+ # We do NOT pass custom_system_prompt to RLM because RLM_SYSTEM_PROMPT
+ # uses Python .format() with {custom_tools_section} placeholder and
+ # double-escaped braces {{chunk}} in examples. Appending to the template
+ # and letting build_rlm_system_prompt() .format() it would break on any
+ # curly braces in the appended text (skills, memory, system message).
+ # Instead, let RLM use its default system prompt (with proper tool
+ # formatting) and inject MachinaOS context into the user prompt.
+ augmented_prompt = prompt
+ if system_message:
+ augmented_prompt = f"## Instructions\n{system_message}\n\n## Task\n{augmented_prompt}"
+ if supplementary_context:
+ augmented_prompt = f"{supplementary_context}\n\n{augmented_prompt}"
+
def _run_rlm():
from rlm import RLM
from rlm.logger import RLMLogger
@@ -134,20 +149,19 @@ class RLMService:
max_budget=float(max_budget) if max_budget else None,
max_timeout=float(max_timeout) if max_timeout else None,
max_tokens=int(max_tokens) if max_tokens else None,
- custom_system_prompt=system_message if system_message else None,
custom_tools=custom_tools if custom_tools else None,
logger=RLMLogger(), verbose=verbose,
)
- return rlm_instance.completion(prompt)
+ return rlm_instance.completion(augmented_prompt)
result = await asyncio.to_thread(_run_rlm)
# === Memory save (same pattern as execute_chat_agent) ===
if memory_data and memory_data.get('node_id'):
- from services.memory_store import save_message
+ from services.memory_store import add_message
session_id = memory_data.get('session_id', 'default')
- save_message(session_id, "user", prompt)
- save_message(session_id, "assistant", result.response)
+ add_message(session_id, "user", prompt)
+ add_message(session_id, "assistant", result.response or "")
execution_time = time.time() - start_time
iterations = len(result.metadata.get("iterations", [])) if result.metadata else 0