Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

continuity-framework

bởi riley-coyote

Notes & PKM
1.3kLượt tải do clawskills.sh công bố
1Sao
0Lượt xem
0 lượt thíchĐăng nhập để thích kỹ năng này
—Chưa có đánh giá
Đánh giá của bạn

Tóm tắt

Asynchronous reflection and memory integration.

Mô tả

Kỹ năng này làm gì

Runs asynchronous reflection after AI sessions end, extracting typed memories with confidence scores. Generates follow-up questions from each session and surfaces them when the user returns. Separate memory types track facts, commitments, preferences, and relationships across conversations.

Unlike passive note logging that requires manual review, it actively processes session content and generates questions the agent genuinely wants to ask rather than performing familiarity.

Khi nào nên dùng

  • Surfacing forgotten follow-up questions at the start of a new session
  • Tracking commitments made across multiple conversations
  • Triggering automatic reflection after conversation idle time
  • Building a persistent self-model for an AI agent over time
  • Reviewing which inferences the agent holds with high vs low confidence

Quy trình ví dụ

ĐẦU VÀO

User asks: 'Anything you wanted to follow up on?'

AGENT
  1. 1

    Detects session ended with idle time exceeding the configured threshold

  2. 2

    Runs continuity reflect to analyze the most recent conversation

  3. 3

    Extracts structured memories with types such as fact, commitment, and relationship, each with a confidence score

  4. 4

    Stores generated questions in questions.md for the next session

  5. 5

    Runs continuity greet to surface the top pending questions to the returning user

ĐẦU RA

A context-appropriate greeting with up to three pending questions drawn from prior session reflection

Cài đặt

clawhub install riley-coyote/continuity-framework

Chạy lệnh CLI, hoặc dán prompt vào Claude Code, Cursor, Codex hay coding agent bạn đang dùng.

Liên kết