@fastagent/cli 0.6.13 → 0.6.16
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +14 -0
- package/cli.js +388 -347
- package/examples/fastagent.config.example.json +1 -0
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -20,6 +20,7 @@
|
|
|
20
20
|
- MCP 工具接入与 MCP 资源读取
|
|
21
21
|
- skills 安装、发现和调用
|
|
22
22
|
- 记忆:从 project / user 记忆目录召回可复用的长期信息,并在稳定回合后沉淀新的记忆
|
|
23
|
+
- Dream: 记忆整理,以及后台 auto-dream 记忆巩固 (支持手动/dream)
|
|
23
24
|
- 上下文压缩:长会话自动压缩历史上下文
|
|
24
25
|
- 定时任务(`cron`):会话级持久化定时调度
|
|
25
26
|
- 计划清单:管理会话内的任务列表(task_**)
|
|
@@ -103,6 +104,18 @@ fastagent doctor --config ./fastagent.config.json
|
|
|
103
104
|
- 如果设置了 `FASTAGENT_SUPPORTS_PREVIOUS_RESPONSE`,值必须是 `true|false|1|0`;非法值会在
|
|
104
105
|
CLI runtime config 解析阶段直接报错
|
|
105
106
|
|
|
107
|
+
模型切换兼容性提示:
|
|
108
|
+
|
|
109
|
+
- `supportsPreviousResponse` 只控制 `openai-responses` provider 是否走 `previous_response_id` replay,不代表任意模型历史都能无损复用。
|
|
110
|
+
- 某些模型族有自己的 continuation 要求。当前已知例子是 `kimi-k2.5`:当 thinking 开启且进入多步 tool loop 时,后续请求需要保留 assistant turn 里的 `reasoning_content`。
|
|
111
|
+
- 因此,同一会话里如果从别的模型族切到 `kimi-k2.5`,而旧历史里又包含工具调用链路,可能会因为缺少 Kimi 所需 continuation 而要求新会话或 `/reset`;纯文本历史通常更容易兼容。
|
|
112
|
+
|
|
113
|
+
默认 Thinking Level:
|
|
114
|
+
|
|
115
|
+
- `engine.defaultThinkingLevel` 和 `FASTAGENT_DEFAULT_THINKING_LEVEL` 用来控制 engine 默认推理强度,合法值是 `off|minimal|low|medium|high|xhigh`
|
|
116
|
+
- CLI 会把这个值同时投影到 ACP runtime options 和 IM env snapshot
|
|
117
|
+
- 如果你准备把默认模型切到 `kimi-k2.5`,又希望尽量避开 thinking continuation 历史要求,可以把默认值设成 `off`
|
|
118
|
+
|
|
106
119
|
示例最小配置(IM 通道运行):
|
|
107
120
|
|
|
108
121
|
```json
|
|
@@ -119,6 +132,7 @@ fastagent doctor --config ./fastagent.config.json
|
|
|
119
132
|
"maxRetries": 2
|
|
120
133
|
},
|
|
121
134
|
"engine": {
|
|
135
|
+
"defaultThinkingLevel": "medium",
|
|
122
136
|
"turnTimeoutMs": 180000,
|
|
123
137
|
"toolExecutionTimeoutMs": 120000
|
|
124
138
|
},
|