yugoo-design 0.1.2 → 0.2.2

package/payload/shared/bundles/research/SKILL.md

@@ -41,7 +41,11 @@ working copy. They never use host filesystem or shell access.
 
 - **Plan.** Maintain the complete plan state in `research/research.json`: rounds
   with directions (`title` + `research_points`) and deterministic direct
-  artifact paths for new work. The Producer updates its working copy and calls
+  artifact paths for new work. Preserve explicit source limits, allowed source
+  types, product scope and exclusions in the affected direction's existing
+  `research_points`, alongside its questions, so the saved plan retains the
+  complete brief across confirmation, delegation and serial takeover. Do not
+  invent a limit when the user supplied none. The Producer updates its working copy and calls
   `design_publish({action:"handoff"})` to enter the independently configured gate,
   and the runtime presents `开始调研` / `Start Research`; the user may revise
   the plan before starting it.
@@ -149,6 +153,34 @@ the current product or requested exploration. Express this in the existing
 `research_points`, not extra plan fields. Broad landscape research can discover
 useful questions; it need not pretend the user has already chosen a solution.
 
+### Honor the user's evidence bounds
+
+The user's source limit is an acceptance condition for the requested output,
+separate from the Host's tool/time/token budget. Carry the complete current
+constraint into every Worker brief. Before serial research or taking over a
+failed Worker, reread the saved direction and apply the same constraint; a failed
+attempt does not permit a larger source set or broader product scope.
+
+For a limit on report sources, count distinct cited document URLs across the
+findings and its `信息来源` section, not domains, repeated citations or search
+result counts. For example, “at most two first-party sources per direction”
+permits at most two distinct first-party document URLs in that direction's
+findings and matching `source_urls`. It does not mean two websites with unlimited
+pages, nor two new sources in addition to retained ones. A limit on pages opened
+or tool calls instead applies before retrieval; follow the user's actual wording.
+
+Before integrating a Worker receipt or publishing serial findings, inspect the
+landed report against the current direction's questions and bounds. If it exceeds
+the source limit, keep the strongest permitted evidence and revise or qualify
+every claim that depended on a removed source. Update inline links, the source
+list and `source_urls` together; trimming only JSON or the bibliography does not
+repair an over-budget report. Regenerate the affected round conclusion from the
+accepted findings without adding external sources. If the permitted evidence
+cannot support the requested answer, keep that limitation explicit rather than
+silently adding a source. Report the actual per-direction source counts and any
+unmet bound in the handoff. Ordinary Bundle validation does not mechanically
+interpret these free-text constraints; passing it never waives this review.
+
 Apply the same evidence standard to every statement, including headings,
 table cells, chart labels and recommendation reasons:
 - An observation says only what the source establishes, within its context.
@@ -399,7 +431,8 @@ For each direction:
 1. Cover every `research_points` bullet from more than one useful search angle.
    Stop when credible retrieved evidence supports a useful answer, or when the
    Host's configured tool/time/token budget is exhausted. The Host owns those
-   limits; this Skill Body does not promise a fixed number of search or fetch
+   runtime limits; the user's evidence/output bounds still apply to both Worker
+   and serial execution. This Skill Body does not promise a fixed number of search or fetch
    calls. If important evidence remains missing at the configured boundary,
    leave the gap explicit instead of extending the run indefinitely.
 2. Open the most authoritative hits with `web_fetch` when available — search
@@ -410,8 +443,9 @@ For each direction:
    writes and repairs that exact path; the Producer reads the file, checks the
    receipt, and updates the direction state.
 
-When delegating a direction, include this bounded-evidence policy, all of that
-direction's `research_points`, and the complete findings contract below in the
+When delegating a direction, include this bounded-evidence policy, the user's
+explicit source/scope constraints, all of that direction's `research_points`,
+and the complete findings contract below in the
 Worker prompt. Search depth must not depend on whether the Producer performs the
 direction itself or delegates it.
 

package/payload/shared/bundles/research/locales/en-US.json

@@ -1,3 +1,3 @@
 {
-  "guidance": "Select a round or research direction to inspect findings and sources. If there is no plan, describe the question in the main conversation.",
+  "guidance": "Check whether the findings answer your question, then whether the key sources support them. Point out weak evidence or directions that need further investigation.",
   "bundleDescription": "- **Capability**: external information gathering — market context, competitor analysis, business background, reference targets (\"what is X\").\n- **Resolves**: U gaps caused by uncertain external facts, reference products, domain conventions, regulations, or current landscape.\n- **Candidate when**: reliable external evidence would materially change a product decision and a confident answer would otherwise require guessing.\n- **Do not infer**: a named product, an unfamiliar word, or a request for a bounded draft does not by itself require a research round. Analysis explicitly requested as input to a later imitation/build does require research.\n- **Output**: `research` panel — sourced findings, their design implications and evidence limits, persisted to design state.","title":"Research","empty":"No research plan yet. Describe the external question to investigate in the conversation."}

package/payload/shared/bundles/research/locales/zh-CN.json

@@ -1,3 +1,3 @@
 {
-  "guidance": "选择轮次或调研方向查看发现和来源；还没有计划时，在主对话说明要核查的问题。",
+  "guidance": "先看结论是否回答了你的问题，再核对关键来源能否支持它；指出证据不足或需要继续调查的方向。",
   "bundleDescription": "- **能力**：收集外部信息，包括市场背景、竞品分析、业务背景和参考对象（“X 是什么”）。\n- **解决**：由不确定的外部事实、参考产品、领域惯例、法规或当前市场格局造成的 U 类缺口。\n- **适用时机**：可靠的外部证据会实质影响产品决策，而直接给出确定答案将依赖猜测。\n- **不要推断**：仅仅出现产品名、生词或要求范围明确的草案，并不需要启动调研；明确要求为后续仿制或构建提供分析时则需要调研。\n- **输出**：包含有来源的发现、设计启示和证据局限并持久化到设计状态的 `research` 面板。","title":"调研分析","empty":"暂无调研计划。请在对话中描述需要核查的外部问题。"}

package/payload/shared/bundles/research/reference/artifact-contract.md

@@ -8,6 +8,37 @@ Canonical state: `research/research.json`. Findings: `research/{round_id}/{direc
 
 Use `artifact_kind: "research_direction"`, round-scoped `artifact_id: "{round_id}/{direction_id}"` and the exact findings file for delegation or serial focus. Conclusion focus uses `research_conclusion`, round ID and the exact conclusion file. Each author writes/repairs its artifact; the Producer reads findings for actual synthesis, updates the corresponding refs and sources, then validates and publishes the integrated round. A reference is accepted only after the file exists. There is no separate research script; use the ordinary Bundle validation.
 
+## User bounds and evidence acceptance
+
+For authorized research or revision, preserve explicit source limits, allowed
+source types, product scope and exclusions in the affected direction's existing
+`research_points`. They accompany the questions in every Worker brief and remain
+binding when the Producer takes over a failed Worker. Host runtime budgets do
+not relax those output constraints, and no source limit is implied when none was
+requested.
+
+A report-source cap counts distinct cited document URLs, including inline
+citations and the `信息来源` list; repeated use of one URL counts once, while two
+pages on one site count as two sources. The final source list and `source_urls`
+must match the report's used sources. A retrieved search-result list is not the
+report's accepted source set. If a figure uses a separate asset URL, retain its
+link to the cited source document rather than treating an unexplained asset as
+new evidence. An explicit limit on pages opened or tool calls applies to retrieval
+instead; follow the user's stated unit and scope.
+
+The Producer checks these bounds against the actual findings before acceptance
+and again before publication, for both delegated and serial output. Repair an
+excess by revising claims and links together; never hide extra evidence by only
+shortening the bibliography or JSON. A conclusion cites only sources retained in
+accepted findings. If the bounded evidence is insufficient, explain the gap
+without silently exceeding the cap. Report actual per-direction source counts
+and unmet constraints. The existing schema and declared rules do not infer a
+numeric limit from free text, so structural validation is not proof of compliance.
+
+Restoring existing research preserves its observed sources and provenance;
+report an apparent historical mismatch instead of deleting evidence to satisfy
+a newly requested limit without an authorized revision.
+
 ## Findings Markdown
 
 User-facing report, not a raw JSON payload:

package/payload/shared/bundles/research/reference/preview.json

@@ -8,7 +8,7 @@
         "D-PREVIEW-001": {
           "direction_id": "D-PREVIEW-001",
           "title": "从保存到再次阅读的使用场景",
-          "research_points": ["了解材料难以再次找到的典型情境。", "比较按标题查找与按主题整理的使用边界。"],
+          "research_points": ["了解材料难以再次找到的典型情境。", "比较按标题查找与按主题整理的使用边界。", "合成约束示例：本方向报告最多引用 2 个不同的一手来源文档；不足之处明确说明。"],
           "findings_file": null
         },
         "D-PREVIEW-002": {

package/payload/shared/bundles/research/skill.yaml

@@ -2,7 +2,7 @@
 
 # Stable internal id; must equal this directory name.
 name: research
-version: "1.0"
+version: "1.1"
 
 # Optional candidate hint, not automatic execution or a unique provider binding.
 review_support: [reflection]

package/payload/shared/bundles/research/validation.yaml

@@ -3,6 +3,9 @@
 # declared below. Citation provenance and report quality are reviewed by the
 # executing Agent against SKILL.md, not by a separate citation/report driver
 # declared in this Bundle. A successful validation is not proof of grounding.
+# Explicit per-direction source/scope bounds in research_points are also
+# Agent-reviewed against the actual report; these rules do not infer numeric
+# source limits from free text. Passing them does not waive the user's bounds.
 #
 # Contract summary:
 # - Every accepted direction has a direct `findings_file` and non-empty file.

package/payload/shared/bundles/role_access/locales/en-US.json

@@ -35,7 +35,7 @@
   "parentRole": "Source role",
   "inheritanceReason": "Evidence for inheritance (optional)",
   "addChildRole": "Add child role",
-  "guidance": "Roles are flat by default. Set a source role only for evidenced permission reuse. Each role has at most one parent and keeps its own assignment rule.",
+  "guidance": "Check what each role can do and which data it can access, looking for excessive or missing permissions. Set one source role only when reuse is needed; grants remain independent.",
   "graphHint": "Connect the child to its source role. Return to details for local permissions and read-only inherited sources.",
   "inheritanceHelp": "Inherited permissions retain their original applicability, confirmation and identity verification. Role assignment and conflict precedence are not inherited."
 }

package/payload/shared/bundles/role_access/locales/zh-CN.json

@@ -35,7 +35,7 @@
   "parentRole": "来源角色",
   "inheritanceReason": "继承依据（可选）",
   "addChildRole": "添加子角色",
-  "guidance": "角色默认平级；只有明确需要复用权限时才设置来源角色。每个角色最多一个来源，授予规则仍独立。",
+  "guidance": "检查各角色能做什么、作用于哪些数据，是否存在权限过多或遗漏；仅在确需复用时指定一个来源角色，授予规则仍独立。",
   "graphHint": "从子角色连向来源角色；返回详情查看本角色权限及只读的继承来源。",
   "inheritanceHelp": "继承权限保留原适用条件与确认、身份验证要求；不会继承角色授予规则，也不自动决定冲突优先级。"
 }

package/payload/shared/bundles/smart_questions/locales/en-US.json

@@ -1,5 +1,5 @@
 {
-  "guidance": "Answer the current round. Explain uncertainty or leave an answer empty; previous rounds are read-only.",
+  "guidance": "Check whether the questions reflect your intent, then answer or correct them. Explain uncertainty or leave it open; earlier rounds remain available to review.",
   "bundleDescription": "- **Capability**: maintain coherent product background and evolving context through editable questions and answers.\n- **Resolves**: understanding of users, needs, value, scenarios and boundaries, plus relevant context not owned by another design facet.\n- **Candidate when**: exploring requirements, reconstructing product background, revising existing understanding or clarifying context necessary for the current result; not a technical inventory or fixed questionnaire.\n- **Do not infer**: a short request, low score, or missing detail alone does not authorize questioning when a useful draft or grounded default is what the user requested.\n- **Output**: `smart_questions` panel — editable question cards with rendered options and answers.",
   "configuration": "Application Development Configuration",
   "empty": "There are no questions to answer.",

package/payload/shared/bundles/smart_questions/locales/zh-CN.json

@@ -1,5 +1,5 @@
 {
-  "guidance": "回答当前轮次的问题，不确定时可说明或留空；历史轮次只供回看。",
+  "guidance": "确认问题中的理解是否符合你的意图，再回答或纠正；不确定的地方可以说明或留空，历史轮次可回看。",
   "bundleDescription": "- **能力**：持续维护连贯的产品背景与关键上下文，并在需要时通过可编辑问题向用户取证。\n- **解决**：产品为谁、在何种场景解决什么问题、带来什么价值及其边界；补充其他设计维度不适合承载但影响整体理解的上下文。\n- **适用时机**：探索需求、重建产品背景、修订已有理解或澄清当前结果的关键缺口；不是技术盘点或固定问卷。\n- **不要推断**：需求简短、清晰度分数较低或细节缺失，并不意味着可以提问；用户需要可用草案或有依据的默认值时应直接推进。\n- **输出**：可编辑问题卡、选项和答案组成的 `smart_questions` 面板。",
   "configuration": "应用开发配置",
   "empty": "当前没有待回答的问题。",

package/payload/shared/bundles/web_app_prd_assembly/locales/en-US.json

@@ -1,5 +1,5 @@
 {
-  "guidance": "Browse the PRD by chapter and use source labels to return to the design. Request generation or revision in the main conversation.",
+  "guidance": "Check that the PRD reflects the agreed scope and key decisions. Follow source labels back to the original design when something is wrong, or request a revision in the conversation.",
   "bundleDescription": "- **Capability**: generate or update one development-ready Web App PRD from current design evidence and user material.\n- **Resolves**: the delivery gap where product, design, engineering, test, and operations still lack one shared implementation-ready specification.\n- **Candidate when**: the user asks to generate, export, update, or inspect a complete PRD.\n- **Do not infer**: PRD assembly is an on-demand delivery action, not a required design gate and not permission to mutate owner Bundles.\n- **Output**: one stable canonical `prd.html`; Git retains history.",
   "contents": "Contents", "empty": "No PRD yet. Ask in the conversation to generate or revise the complete PRD.",
   "pending": "Pending confirmation", "legend": "Purple rail: temporary completion to keep the PRD coherent; not yet confirmed by the user",

package/payload/shared/bundles/web_app_prd_assembly/locales/zh-CN.json

@@ -1,5 +1,5 @@
 {
-  "guidance": "按目录查看 PRD，并通过来源标签返回原设计；生成或修订请在主对话提出。",
+  "guidance": "核对 PRD 是否准确表达已确认的范围和关键决策；发现偏差时通过来源标签回到原设计，或在对话中提出修订。",
   "bundleDescription": "- **能力**：从当前设计证据和用户材料生成或修改一份可直接用于研发的完整 Web App PRD。\n- **解决**：产品仍缺少一份各交付角色都能直接理解、审阅和实施的统一需求文档。\n- **适用时机**：用户要求生成、导出、修改或查看完整 PRD。\n- **不要推断**：PRD 整合是按需交付动作，不是必经设计关卡，也不允许修改证据所属 Bundle。\n- **输出**：唯一稳定的 canonical `prd.html`；Git 保留历史修订。",
   "contents": "文档目录", "empty": "暂无 PRD。可以在对话中要求生成或修订完整 PRD。",
   "pending": "待确认", "legend": "紫色左轨：为保持 PRD 结构完整而临时补全，尚待用户确认",

package/payload/shared/bundles/whiteboard/locales/en-US.json

@@ -1,3 +1,3 @@
 {
-  "guidance": "Select or add a topic and express ideas with text, shapes and connectors. Save after editing.",
+  "guidance": "Check whether the ideas and relationships in the drawing are clear. Add text, shapes or connections where something is missing; create another topic when the questions are distinct.",
   "bundleDescription": "Draw and revise topic-based diagrams together using text, shapes and connectors; explore ideas without replacing formal flows or page designs.","empty":"Create a topic to start drawing together.","boards":"Whiteboards","board":"Topic","id":"ID","title":"Topic title","file":"Scene file","new":"New topic"}

package/payload/shared/bundles/whiteboard/locales/zh-CN.json

@@ -1,3 +1,3 @@
 {
-  "guidance": "选择或新增主题，用文字、图形和连线表达想法；编辑后保存。",
+  "guidance": "检查图中的想法和关系是否表达清楚，用文字、图形或连线补上遗漏；需要区分不同问题时可新增主题。",
   "bundleDescription": "按主题共同绘制和修改文字、图形与连线；用于开放探索，不替代业务流程或页面设计的正式产物。","empty":"新建一个主题，开始共同绘制。","boards":"自由白板","board":"主题","id":"标识","title":"主题名称","file":"场景文件","new":"新主题"}

package/payload/shared/kits/web_app/kit.yaml

@@ -14,7 +14,7 @@ mounts:
   # Research markets, competitors and domain context, with sources and design implications.
   - bindingKey: research
     name: research
-    version: "1.0"
+    version: "1.1"
 
   # 功能模块：将产品目标与使用场景展开为可逐项推敲的功能卡片。
   # Turn product goals and scenarios into function cards that can be reviewed individually.

package/payload/shared/kits/web_app/mindset.md

@@ -1,6 +1,6 @@
 ---
 name: web_app
-version: "1.1"
+version: "1.2"
 description: >
   Calibration guidance for design choices in general web applications, SaaS,
   and enterprise products. It ranks genuine alternatives; it never chooses a

package/payload/shared/knowledge/runtime-manifest.json

@@ -5,14 +5,16 @@
     "@ladybugdb/wasm-core": "0.21.1",
     "@huggingface/transformers": "3.8.1",
     "onnxruntime-common": "1.21.0",
-    "undici": "8.10.2"
+    "undici": "8.10.2",
+    "trash": "10.1.1",
+    "yaml": "2.9.1"
   },
   "assets": {
     "cli-lifecycle.mjs": "039c7705557b58c435c0a96a6540b2a233a3583ebab36c5d100bad0771a15d88",
     "embedding-worker.js": "1dc5f99a02fc8bd58ab318991a2145665c95bc65f4a025b63be524dffdb66f00",
     "embedding-worker.mjs": "1dc5f99a02fc8bd58ab318991a2145665c95bc65f4a025b63be524dffdb66f00",
     "package.json": "a3e820bf1715afda0b535dd848a22362dbdc28296fbbbe6ffd6b7712d06194b6",
-    "runtime.mjs": "08e99eae9738441b5a1a588a5a1a16a986e0dd8a8a41199fc88d19ab0956940e"
+    "runtime.mjs": "1021590ab75d4eeb8f00b7fbc61f68c15869c755432b83f61d5d45c7131ac3d9"
   },
   "officialInputs": {
     "repository": "https://github.com/deepseek-ai/deepseek-harness",

package/payload/shared/knowledge/runtime.mjs

@@ -6894,7 +6894,7 @@ var require_public_api = /* @__PURE__ */ __commonJSMin(((exports) => {
 	exports.visitAsync = visit.visitAsync;
 })))();
 //#endregion
-//#region packages/core/src/runtime-drivers.ts
+//#region packages/core/src/design-driver-id.ts
 const DRIVER_ID = /^[a-z][a-z0-9]*(?:[.-][a-z0-9]+)*(?:\/v[1-9][0-9]*)$/;
 /**
 * Whether a string is a well-formed versioned dotted identifier ending in `/vN`.
@@ -6905,16 +6905,6 @@ function isValidVersionedId(value) {
 	return DRIVER_ID.test(value);
 }
 //#endregion
-//#region packages/core/src/validation-rules.ts
-/**
-* validation.yaml rule engine. Evaluates a skill's declarative validation
-* contract (issue codes, messages, repair hints) in one of three scopes:
-* `state` (checkpoint on the persisted JSON), `completion` (terminal
-* deliverables, including disk-backed artifacts), and `candidate` (child-result
-* gates applied before Producer integration). Declarative primitives cover the common
-* shapes; Bundle-provided drivers carry procedural validation.
-*/
-//#endregion
 //#region packages/core/src/renderer.ts
 /** Generic Design-state rendering for declarative and runtime-contributed Bundles. */
 //#endregion
@@ -6924,30 +6914,18 @@ const PANEL_COMPONENT_CATALOG = {
 		description: "Switch between uniquely identified, titled children with keyboard-accessible tabs; inactive content stays mounted to preserve drafts.",
 		autoOrder: 2
 	},
-	split: {
-		description: "Compose sidebar and main slots.",
-		autoOrder: 3
-	},
 	stack: {
 		description: "Arrange child components vertically.",
 		autoOrder: 1
 	},
 	panel_workspace: {
-		description: "Full-height workspace with a resizable navigation sidebar.",
+		description: "Layout shell with sidebar/main slots. Plain main content scrolls within the shell; use record_workspace with record_inspector for record selection and editing.",
 		ownsWorkspace: true
 	},
 	text_block: {
 		description: "Show escaped plain-text guidance; never executable markup.",
 		autoOrder: 4
 	},
-	record_nav: {
-		description: "Navigate a declared record collection.",
-		sourceBound: true
-	},
-	record_tree: {
-		description: "Show record hierarchy using declared parent bindings.",
-		sourceBound: true
-	},
 	record_form: {
 		description: "Edit records using declared fields and schema constraints. Static sections may opt into form.header detail/card, form.reorder:false and finite form.body field/row/group/tabs composition. Record-local tabs reuse mounted shared tabs; titled collapsible groups preserve their values. Unconsumed visible fields remain rendered. Fields can opt into choice_layout, radio_custom user-triggered custom_focus, select allow_clear:false and token_list token_add_mode:menu for presets plus a cancellable custom-entry card. Omitted configuration preserves legacy behavior; generated-origin forms reject these static extensions.",
 		sourceBound: true,
@@ -7037,12 +7015,17 @@ const PANEL_COMPONENT_CATALOG = {
 		ownsWorkspace: true
 	},
 	record_inspector: { description: "Edit the selected record inside its owning record workspace. Optional authored sources compose same-record root and nested section views, preserving their canonical field bindings." },
+	math_document: {
+		description: "Read mathematical statements beside their plain-language explanations, notation and source pointers. Authored math bindings use declared text/rows fields in a record workspace; optional LaTeX editing uses the existing field/save authority. Rendering is not mathematical verification.",
+		sourceBound: true
+	},
 	record_preview: { description: "Preview the selected record title, body and optional action without executing it. Static form.body groups may embed the same component via sidecar; bindings are declared same-section text fields and the preview follows that exact form record, stacking below fields in narrow containers. Generated-origin forms reject this static extension." },
 	record_related_editor: { description: "Edit child records referencing the selected parent; additions bind that parent explicitly. Optional authored record_table groups the same declared editors into complete table columns without converting canonical records. Optional inherit_from_graph presents ancestors' associated records as read-only source-linked groups, never copies or resolves domain policy conflicts." },
 	relation_graph: { description: "Interact with record nodes and binary edges or ordered role-bearing participants; multi-party relation centers project the original relation record." },
 	reference_graph: { description: "Read-only relation canvas derived from declared in-workspace record references; details edit the original records, never a graph copy. Nodes may use label_scope: record for a root text label declared on the same canonical collection. Links use target or finite targets matching the field options_from_section/options_from_sections; missing or ambiguous identities stay diagnostic without guessed edges." },
+	data_table: { description: "Authored editable sample-data grid with saved column definitions, field types and row identities. Bound to a declared data_table field inside a record workspace; edits use existing save ports. Supports text, number, boolean, date and select cells, field properties and two-index matrix viewing. No file upload or expressions; not enabled for generated panels." },
 	record_matrix: { description: "Pivot a declared rows field inside the selected record." },
-	record_artifact: { description: "Preview an existing artifact of the selected record." },
+	record_artifact: { description: "Preview an existing artifact of the selected record, or inline related artifacts through an explicit record_select reference_field. Missing related records create no empty view; unreadable referenced files remain visible." },
 	structured_data: {
 		description: "Read structured saved data without changing it.",
 		autoOrder: 10
@@ -7741,6 +7724,16 @@ Object.freeze({
 	maxSnapshotBytes: 8388608
 });
 //#endregion
+//#region packages/core/src/panel-save.ts
+/** Shared main-conversation contract for the user's Save and assess action. */
+const DESIGN_PANEL_SAVE_ASSESSMENT_GUIDANCE = [
+	"The user chose Save and assess next step. After a meaningful Panel save, you must assess the latest saved Design evidence against the agreed user scope and call report_design_clarity in this same main-conversation continuation before asking the next question, recommending another Bundle or ending your reply.",
+	"Use the current server-supplied project, saved checkpoint and Kit input. If those inputs have changed, read the current Design context before reporting; do not assess a superseded checkpoint as current. Explain known evidence and gaps without inventing facts or a zero score.",
+	"Several saves delivered together need one assessment of their latest relevant state. An unchanged or replayed save does not require another score; reuse an existing assessment only when its saved Design evidence, Kit content and reasoning still apply to the current agreed scope.",
+	"A later user stop, cancellation or instruction not to assess takes priority. If the reporting capability is unavailable or reporting fails, state that limitation and preserve the saved result; a missing score does not block independently authorized Design work.",
+	"No second evaluator is required. Do not start another model or Bundle, reopen a settled invocation, or rewrite saved Design Files just to assess them. Existing user-input handoffs and main-Agent permissions remain in force."
+].join(" ");
+//#endregion
 //#region packages/core/src/validate-design.ts
 /**
 * Renderability checkpoint (the second validation layer). Given a design skill
@@ -8219,7 +8212,7 @@ function queryKnowledgeSnapshot(snapshot, request) {
 		}) }
 	});
 }
-`${RECONSTRUCTION_METHOD}`;
+`${RECONSTRUCTION_METHOD}${DESIGN_PANEL_SAVE_ASSESSMENT_GUIDANCE}`;
 //#endregion
 //#region packages/core/src/design-clarity-schema.ts
 /** Same compact model output schema for native and DSH clarity evaluators. */