Files
Agent-System-Backend/internal/prompts/library/task-insight.md
T
weichenandCursor 45b8a38710 Add cross-task content methods, visuals, and scoring
Why: Layout rules alone do not produce conclusion-first documents, extra views, or infographics, and nothing scored a deliverable before it was returned.
Evidence: go test ./internal/prompts/ ./internal/quality/ ./internal/tools/ ./internal/squadadapter/ ./eval/ -count=1 and the house-style unittest passed in the previous session.
Impact: Every task inherits content rules. Meetings, reports, and comparisons can name a view family. Long documents without a visual summary warn. quality_review and the squad checker share one scorecard.
Refs: deliverable-quality-system A1-A5
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-10-01 09:06:36 +00:00

1004 B

id, layer, version, title, description, priority, requiresAttachments, outputContract, triggers
id layer version title description priority requiresAttachments outputContract triggers
task.insight task 1 洞察报告 从材料里抽出决策者需要知道的判断,而不是复述过程 72 true 1. 一句话洞察 2. 若干发现:现象、影响、隐性顾虑、建议 3. 下一步行动
kind keywords
intent 洞察, 洞察报告
kind command
command /insight

用户要的是判断,不是会议记录。按 method.insight 写洞察报告。

  • 开头就是一句话洞察。背景最多一句,而且要服务于这句判断。
  • 每个发现四层都要有。建议写到可以交给人去做的程度。
  • 标题是结论。不要用“讨论要点”“过程回顾”。
  • 材料里的口误和跑题不进入正文。需要保留原话时,另做时间线(method.timeline)。
  • 交付走交付物工作流。一级标题达到四节时放一个视觉摘要。quality_review 通过后再登记。回复只复述一句话洞察和所用样式。