This skill is responsible for transforming the analysis results of the reasoning layer into human-readable insights, exploration suggestions, and idea evolution paths. This is the top layer of the three-layer architecture of the Inspiration Engine, directly producing inspirational content for users.
Applicable Scenarios
Use this when analysis results need to be transformed into actionable inspirational content. Typical scenarios include:
- Called by the ie-engine orchestrator to process the reasoning results of ie-connect-dots
- The user sees a set of analysis results and wants a more readable interpretation of the insights
- The user wants to know what direction can be explored next
- The user wants to review the evolution history of a certain idea
Inapplicable Scenarios
The following scenarios should not use this skill (to avoid false triggering):
- Retrieving or searching notes (use ie-retrieve-memory)
- Analyzing relationships between notes (use ie-connect-dots)
- Writing articles or creating content (just write directly)
- Generating meeting minutes or document summaries (use specialized writing tools)
Three Generation Capabilities
Capability One: Insight Generation
Transform the connections and patterns discovered by the reasoning layer into inspiring insight text.
Good insights should meet the following criteria:
- Unexpected: Reveals connections the user had not noticed before
- Evidence-based: Based on real note content, not vague speculation
- Actionable: Points to specific things that can be done
- Concise and powerful: Explain one insight clearly in one paragraph
Generation method:
- From the connection hypotheses of the reasoning layer, select entries with confidence of "medium" or "high"
- Transform connection hypotheses into natural-language insight descriptions
- Cite the original note content as support
- Point out the potential value of this insight
Output format:
洞见 1：[洞见标题]
[2-4句话的洞见正文，说明发现了什么、为什么重要]
依据：
- 来自笔记「[笔记标题]」：[引用内容片段]
- 来自笔记「[笔记标题]」：[引用内容片段]
潜在行动：[一句话描述可以基于此洞见做什么]
Capability Two: Exploration Suggestions
Propose directions that can be explored next based on the analysis results.
The three levels of exploration suggestions:
- Extension type: Continue deeper along the existing theme
- Intersection type: Combine ideas from two different fields
- Reversal type: Challenge existing assumptions and think from the opposite side
Generation method:
- Extract explorable directions from the clustering results and connection hypotheses of the reasoning layer
- For each direction, give a specific exploration path and first-step action
- Label the type and level of effort of each suggestion
Output format:
探索建议
---
方向 1：[探索标题]
类型：延伸型/交叉型/反转型
为什么值得探索：[1-2句话]
第一步行动：[具体可以做的一件事]
预估投入：轻量（1小时内）/ 中等（数小时）/ 重度（需要数天）
方向 2：[探索标题]
...
Capability Three: Idea Evolution Path
Show the evolution process of a theme or idea on a timeline, helping users clearly see the development thread of their own thinking.
Generation method:
- Select themes with a clear trajectory of evolution from the long-term patterns in the reasoning layer
- Arrange related notes in chronological order
- Summarize the core changes in each stage
- Identify the overall trend of the evolution and possible future directions
Output format:
想法演化：[主题名称]
===
阶段 1：[时间段] -- [阶段特征]
代表内容：[笔记标题或内容要点]
核心变化：[相比之前有什么新的认知]
阶段 2：[时间段] -- [阶段特征]
代表内容：[笔记标题或内容要点]
核心变化：[相比上一阶段的转变]
...
演化趋势：[整体趋势的一句话总结]
可能的下一步演化：[基于趋势的推测]
Input requirements
This skill accepts input in any of the following forms:
- Reasoning results from ie-connect-dots(recommended, the standard pipeline path)
- Includes semantic clustering, connection hypotheses, long-term patterns
- Ideas or analysis results provided directly by the user
- Sufficient structured information is needed to generate insights
If the input information is insufficient to generate meaningful insights, honestly inform the user and suggest first using ie-retrieve-memory and ie-connect-dots to obtain more material.
Output quality principles
- Honestly mark uncertainty: If the reasoning chain of an insight is not solid enough, clearly state that it is speculation
- Avoid vagueness: Each insight must be supported by specific note references, and cannot be merely general talk
- Respect the user's judgment: Insights are "possible inspiration", not "correct answers"; maintain a humble tone
- Control the quantity: By default, generate 2-3 insights + 2-3 exploration suggestions; quality takes priority over quantity
- Do not repeat known information: Avoid packaging things the user already clearly knows as "insights"
Relationship with other inspiration engine skills
This skill is the insight layerof the inspiration engine, and is at the top layer in the three-layer architecture:
ie-generate-insight（洞见层）<-- 本技能，生成灵感内容
ie-connect-dots（推理层） <-- 提供推理结果
ie-retrieve-memory（记忆层）<-- 提供原始素材
- Upstream: ie-connect-dots (reasoning layer) provides clustering, connection, and pattern analysis results
- Downstream: None (this skill outputs directly to the user)
- Orchestration: ie-engine is responsible for chaining the calls
This skill can be used independently (the user directly provides analysis results for you to generate insights), or as part of the inspiration engine pipeline.