feat(service-ai): open framework AI is query-only, declines app-building#1803
Merged
Conversation
test(e2e): lookup quick-filter coverage (CRM parity) (#1682) objectui@76c38108544dd92abf965c0e07b0e4d92264708a
The unified `data_chat` persona (ADR-0040) advertises that it can BUILD or CHANGE the application, but that capability is supplied entirely by the cloud AI Studio plugin's `metadata_authoring`/`solution_design` skills and their tools. On the open single-env framework those skills are not registered, so the authoring tools never resolve — yet the LLM, still reading the "you can build" persona, role-plays designing a whole system (emitting design docs it has no tools to execute). Users saw the standalone open framework "develop" a full 进销存 app that it could not actually create. Fix: in buildSystemMessages, when no authoring (build-register) skill is active, append a deployment-capability note constraining the assistant to data/query and instructing it to decline build requests instead of pretending. Keyed off actual skill presence, so cloud/EE (AI Studio loaded) keeps the full build UX with zero extra wiring. Adds tests for both the absent and present build-register cases. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
问题
用户单独运行开源 framework（单环境运行时）时，
_console/ai的助手会把自己当成「开发助手」——给它发「帮我开发一个进销存」，它会输出整套系统设计方案、模块表、对象清单，表现得像在「开发应用」。但开源 framework 并不包含 AI 应用搭建能力（那是云端service-ai-studio插件提供的metadata_authoring/solution_designskills 及其工具）。所以它只是在「演」开发——实际一个对象都建不出来，对开源用户是误导。按产品边界：开源 framework 的 AI 只应支持数据查询，不应支持开发（开发是付费/云端能力）。
根因
ADR-0040 把助手统一成单一
data_chatpersona，其 instructions 宣称「You can both ANSWER QUESTIONS ... and BUILD or CHANGE the application itself」。authoring 工具走的是「云端 AI Studio 插件挂载时才注册」的优雅降级——但只降级了工具，没降级人设。开源部署里工具不存在，LLM 仍照着「我能搭建」的人设侃侃而谈。改动
buildSystemMessages：当本次激活的 skills 里没有 authoring（build-register）skill（metadata_authoring/solution_design）时，追加一段「本部署能力说明」——明确「应用搭建/开发不可用，你只能查询/聚合/搜索既有数据并执行应用已暴露的 action；用户要你搭建/开发应用时，不要设计、不要假装，简短说明本版本不提供 AI 应用搭建，转而帮其探索数据」。autoPublishAiBuilds）。测试
chatbot-features.test.ts全绿（64/64，含新增 2 条）。service-ai 套件 256 个断言通过；2 个测试文件因新建 worktree 里@objectstack/types/@objectstack/spec尚未构建而无法 import（与本改动无关的构建顺序问题，CI 构建全图后即通过）。🤖 Generated with Claude Code