refine agent loop and working memory tools

2026-02-10 13:15:31 +08:00
parent 3a9a5ee2aa
commit 8f593c4eaf
3 changed files with 22 additions and 18 deletions
--- a/agent_loop.py
+++ b/agent_loop.py
@@ -73,5 +73,9 @@ def agent_runner_loop(client, system_prompt, user_input, handler, tools_schema,
            datastr = json.dumps(outcome.data, ensure_ascii=False, default=json_default) if type(outcome.data) in [dict, list] else str(outcome.data) 
            next_prompt += f"<tool_result>\n{datastr}\n</tool_result>\n\n"
        next_prompt += outcome.next_prompt
+        if (turn+1) % 5 == 0:
+            next_prompt += f"\n\n[DANGER] 已连续执行第 {turn+1} 轮。禁止无效重试。若无有效进展，必须切换策略：1. 探测物理边界 2. 请求用户协助。"
+        if turn == 23:
+            next_prompt += f"\n\n[DANGER] 已连续执行第 {turn+1} 轮。你必须总结情况进行ask_user，不允许继续重试。"
        messages = [{"role": "user", "content": next_prompt}]
    return {'result': 'MAX_TURNS_EXCEEDED'}