Bug Description
Description:
The agent’s tool call output consistently freezes at the [Calling tool: tool_name with arguments: {...}] marker. The response never completes — no tool result, no follow-up message, no error. The agent has made the call but the response never arrives back.
The command /goal reliably unblocks the frozen state and the agent resumes producing output within 2-5 seconds.
Steps to Reproduce
Steps to reproduce the behavior:
Have a normal conversation with the agent (any model, any provider).
Stop sending messages for 2-3 minutes.
Send a new request that requires tool calls (e.g., “read this file”, “run this command”).
Observe: the response stops at [Calling tool: tool_name with arguments: {...}] and never completes.
Send /goal — the agent resumes within seconds.
The issue is not sporadic — during active back-and-forth conversation, tool calls work fine. It is specifically the first tool call after a pause that hangs.
Expected Behavior
Tool call results should be returned and displayed regardless of idle time between messages.
Actual Behavior
the symptom is purely textual: the conversation wedges at the [Calling tool: ...] marker and produces no further output until /goal is sent. Such as,
[Calling tool: execute_code with arguments={“code”: “from hermes_tools import read_file\n\n# Try reading in chunks\nfor offset in range(1, 1000, 200):\n r = read_file(path="/Users/eb/.hermes/config.yaml", offset=offset, limit=200)\n print(f"— offset {offset} —")\n print(r["content"])\n if "truncated" not in r or not r["truncated"]:\n break”}]
Affected Component
CLI (interactive chat)
Messaging Platform (if gateway-related)
N/A (CLI only)
Debug Report
Debug report uploaded:
Report https://paste.rs/21p96
agent.log https://paste.rs/j2IzB
gateway.log https://paste.rs/OADPl
Operating System
macOS (26.3.1)
Python Version
3.9.6
Hermes Version
0.14.0
Additional Logs / Traceback (optional)
Root Cause Analysis (optional)
No response
Proposed Fix (optional)
No response
Are you willing to submit a PR for this?
Bug Description
Description:
The agent’s tool call output consistently freezes at the [Calling tool: tool_name with arguments: {...}] marker. The response never completes — no tool result, no follow-up message, no error. The agent has made the call but the response never arrives back.
The command /goal reliably unblocks the frozen state and the agent resumes producing output within 2-5 seconds.
Steps to Reproduce
Steps to reproduce the behavior:
Have a normal conversation with the agent (any model, any provider).
Stop sending messages for 2-3 minutes.
Send a new request that requires tool calls (e.g., “read this file”, “run this command”).
Observe: the response stops at [Calling tool: tool_name with arguments: {...}] and never completes.
Send /goal — the agent resumes within seconds.
The issue is not sporadic — during active back-and-forth conversation, tool calls work fine. It is specifically the first tool call after a pause that hangs.
Expected Behavior
Tool call results should be returned and displayed regardless of idle time between messages.
Actual Behavior
the symptom is purely textual: the conversation wedges at the [Calling tool: ...] marker and produces no further output until /goal is sent. Such as,
[Calling tool: execute_code with arguments={“code”: “from hermes_tools import read_file\n\n# Try reading in chunks\nfor offset in range(1, 1000, 200):\n r = read_file(path="/Users/eb/.hermes/config.yaml", offset=offset, limit=200)\n print(f"— offset {offset} —")\n print(r["content"])\n if "truncated" not in r or not r["truncated"]:\n break”}]
Affected Component
CLI (interactive chat)
Messaging Platform (if gateway-related)
N/A (CLI only)
Debug Report
Operating System
macOS (26.3.1)
Python Version
3.9.6
Hermes Version
0.14.0
Additional Logs / Traceback (optional)
Root Cause Analysis (optional)
No response
Proposed Fix (optional)
No response
Are you willing to submit a PR for this?