From e4f8d45d17de818442efdccdb3691be4e38a520d Mon Sep 17 00:00:00 2001 From: DemonGatanjieu <207458273+DemonGatanjieu@users.noreply.github.com> Date: Fri, 30 Oct 2026 03:48:12 +0800 Subject: [PATCH] Guide flexible prompt poses through MCP --- CHANGELOG.md | 3 +++ api/mcp_actions.py | 8 ++++++-- docs/architecture/backend.md | 4 ++++ docs/guides/mcp.md | 2 +- 4 files changed, 14 insertions(+), 3 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index eb03340..eb4c858 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -2,6 +2,9 @@ ## v1.63.0 — unreleased +- AI 写出图提示词时，AMB 的工具说明会建议：姿势可以选时优先选模型容易画清楚的动作，避免不必要的复杂手势；你指定的姿势和提供的文本照你的要求处理。复杂姿势可以用参考姿势或分步修复，并检查结果。 +- When an AI composes an image prompt, AMB's tool descriptions suggest a clear pose suited to the model when the pose is flexible. Explicit poses and supplied text are preserved; pose references or staged repair can help with difficult poses. + - 跑图前的尺寸提醒放宽一档：6GB 显卡采样到约 420 万像素才提醒（以前 210 万），12GB 约 840 万。今天在 6GB 的笔记本显卡上实测，400 万像素一次约 100 秒，并不吃紧；以前的门槛会让 AI 动不动就问你要不要停。解码大图的提醒不变。 - The size warning before a run is less eager: on a 6 GB card from about 4.2 megapixels (2.1 before), 8.4 on 12 GB. A 6 GB laptop card sampled 4 megapixels in about 100 s today; the old line had the AI ask you to stop runs that were fine. The warning for decoding a large picture is as before. diff --git a/api/mcp_actions.py b/api/mcp_actions.py index 5320efd..0ccad63 100644 --- a/api/mcp_actions.py +++ b/api/mcp_actions.py @@ -257,6 +257,10 @@ def speak(ctx, character, text, seed=None, language="auto"): _NODE = {"type": "integer", "description": "A node id."} _MODEL = {"type": "string", "description": "model_id from search_models."} +_PROMPT = {"type": "string", "description": "When composing a prompt, if the pose is flexible, prefer a clear, " + "natural pose suited to the model; avoid unnecessary interlocked hands or complex contact. Preserve " + "explicit user requests and supplied text. For difficult requested poses, consider pose references " + "or staged repair and inspect the result."} def _action(name, title, description, run, properties=None, required=(), destructive=False, open_world=False, @@ -278,7 +282,7 @@ ACTIONS = dict([ _action("set_prompt", "Write a prompt", "Writes text into the workflow's positive or negative prompt box (or the one on node_id). " "mode=append adds to what is there. One Ctrl+Z step.", - set_prompt, {"prompt": {"type": "string"}, "role": {"type": "string", "enum": ["positive", "negative"], "default": "positive"}, + set_prompt, {"prompt": _PROMPT, "role": {"type": "string", "enum": ["positive", "negative"], "default": "positive"}, "node_id": _NODE, "mode": {"type": "string", "enum": ["replace", "append"], "default": "replace"}}, ["prompt"], destructive=True), _action("set_model", "Swap a node's model", @@ -332,7 +336,7 @@ ACTIONS = dict([ "wait_seconds": {"type": "number", "minimum": 0, "maximum": 55, "default": 55, "description": "Seconds to wait for it to end."}, "workflow": {"type": "string", "description": "A name from list_workflows."}, - "prompt": {"type": "string"}, "negative": {"type": "string"}, + "prompt": _PROMPT, "negative": {"type": "string"}, "seed": {"type": "integer", "minimum": 0}, "width": {"type": "integer", "minimum": 16, "maximum": 16384}, "height": {"type": "integer", "minimum": 16, "maximum": 16384}, diff --git a/docs/architecture/backend.md b/docs/architecture/backend.md index 969b60b..ab81e23 100644 --- a/docs/architecture/backend.md +++ b/docs/architecture/backend.md @@ -252,6 +252,10 @@ descriptions. Its picks go to `/anomalous/scan_all` as `targets`, like a list pa ## MCP endpoint +The `prompt` schemas for `set_prompt` and `run_workflow` share concise pose guidance for +agents composing prompts: choose a model-appropriate pose when flexible and preserve explicit +requests and supplied text. This is advice in the MCP tool description; AMB passes text through. + `track_run(prompt_ids, task, wait_seconds)` attaches existing submissions from this server to an open task without running anything. IDs must exist in its queue/history; each is owned by one recorded run. Retrying an identical batch returns that run, not the task's newest run. diff --git a/docs/guides/mcp.md b/docs/guides/mcp.md index 8d6ce90..a125100 100644 --- a/docs/guides/mcp.md +++ b/docs/guides/mcp.md @@ -103,7 +103,7 @@ claude mcp add --transport http anomalous http://127.0.0.1:8188/anomalous/mcp | find_nodes | 搜索最多五个节点摘要；需要具体参数和选项时，用 `types=[节点类型]` 读取 | | add_node / connect_nodes | 添加节点并按唯一类型自动连线，或把节点插进已有路径；见下方例子 | | add_nodes | 一次加一组节点：按 ComfyUI 的 API 格式写，组里的节点互相连，也能连画布上已有的节点（写 `"#12"`，输出写编号或名字），排成一块放在空位上，整组算一步撤销 | -| set_prompt | 改写或追加正向/反向提示词 | +| set_prompt | 改写或追加正向/反向提示词。AI 自己写提示词时，姿势可以选择的话会优先考虑模型容易画清楚的动作；你指定的姿势和提供的文本仍按你的要求处理。`run_workflow` 的提示词也有同样的建议 | | set_model | 换某个加载器节点里的模型 | | add_lora | 插入一个 LoRA，默认接在主模型加载器和已有的 LoRA 后面。LoRA 文件里练过文本编码器的（SD1.5、SDXL、Illustrious 的大多是），CLIP 线也接上：从 checkpoint 接，或者从 UNet 工作流里单独的 CLIP 加载器接，正反向提示词一起改接到它后面；没练过的（多数 Flux、Anima 的 LoRA）只接模型线。LoRA 和主模型的底模对不上时（SD 1.5 的 LoRA 用在 SDXL 上），回复里会提醒 AI 并告诉它按底模去找；换模型（set_model）也一样 | | place_combo | 把一个搭配作为一组新节点摆到画布上，原有节点不动 |