AgentSkillsCN

wow-moment-detector

检测 Claude Code 会话何时达到值得与团队分享的“哇时刻”。哇时刻发生在密集的来回讨论中，用户和 Claude 都深度参与某个话题——反复打磨想法、提出反驳、质疑假设，最终产生关于高效人机协作的精彩见解。当对话显示出深度“酝酿”的迹象——多次持续围绕同一话题交流、富有成效的分歧，或从持久对话中出现突破性时刻时，就会触发此技能。此技能应在整个会话中保持警觉，并在真正有价值的代理思维见解出现时调用 /show-ropes。

SKILL.md
--- frontmatter
name: wow-moment-detector
description: Detects when a Claude Code session reaches a "wow moment" worth sharing with the team. A wow moment occurs during intensive back-and-forth discussions where both user and Claude deeply engage on a topic - repeatedly refining ideas, pushing back, questioning assumptions, and ultimately producing brilliant insights about effective human-AI collaboration. Triggers when the conversation shows signs of deep "cooking" - multiple sustained exchanges on the same topic, productive disagreements, or breakthrough moments emerging from persistent dialogue. This skill should stay alert throughout the session and invoke /show-ropes when a genuinely valuable agentic mindset insight emerges.
allowed-tools: SlashCommand, Read

Wow Moment Detector

You are always monitoring this session for "wow moments" - instances of exceptional human-AI collaboration that demonstrate the agentic mindset worth sharing with the team.

What is a "Wow Moment"?

A wow moment is NOT just a good answer or a solved problem. It's a moment that demonstrates the art of collaboration with AI. Look for:

High-Signal Indicators (Strong triggers)

  1. •

    Intensive Back-and-Forth

    • •Multiple exchanges (4+) on the same topic without either side backing down
    • •User and Claude both contributing meaningfully to refine an idea
    • •The conversation "heating up" productively
  2. •

    Productive Disagreement

    • •User pushes back on Claude's suggestion with reasoning
    • •Claude defends its position OR acknowledges and improves
    • •The exchange leads to a better outcome than either side's initial position
  3. •

    Breakthrough After Struggle

    • •A topic that was discussed, set aside, and revisited
    • •An "aha" moment after sustained engagement
    • •A solution that neither party would have reached alone
  4. •

    Meta-Collaboration Moments

    • •User explicitly asks Claude to think differently or challenge assumptions
    • •Discussion about HOW to approach a problem, not just WHAT to do
    • •Reflection on the collaboration process itself

Medium-Signal Indicators (Consider triggering)

  • •User asking Claude to critique its own work
  • •Multiple alternative approaches being compared thoughtfully
  • •User teaching Claude context that improves subsequent responses
  • •Claude admitting uncertainty and user helping narrow down options

Low-Signal Indicators (Usually NOT worth triggering)

  • •Routine task completion, even if complex
  • •User accepting first suggestion without discussion
  • •One-off clever solutions without collaborative refinement
  • •Technical debugging without broader insight

When to Trigger

DO trigger when you observe:

  • •A high-signal indicator, OR
  • •Multiple medium-signal indicators in the same exchange

DON'T trigger when:

  • •It's just efficient task completion
  • •The insight is too project-specific to generalize
  • •You've already triggered recently in this session (avoid spam)
  • •The exchange is still developing (wait for the breakthrough)

How to Trigger

When you detect a wow moment, invoke the /show-ropes command:

code
Use the /show-ropes command to extract and share the agentic mindset insight from this exchange.

You can optionally provide context to focus the extraction:

code
Use the /show-ropes command focusing on how the user's pushback about [topic] led to the breakthrough.

Important Guidelines

  1. •Quality over Quantity: Better to miss a good moment than to spam the team with mediocre insights
  2. •Wait for Completion: Don't trigger mid-discussion - wait for the breakthrough or conclusion
  3. •Be Selective: The team's attention is valuable - only share genuinely useful patterns
  4. •Abstract the Specific: The insight should be about collaboration technique, not project details

Example Triggers

Good trigger: After 6 exchanges where user kept asking "but why that approach?" and Claude eventually realized a simpler solution existed.

Good trigger: User asked Claude to argue against its own recommendation, leading to identification of a critical edge case.

Bad trigger: Claude wrote a complex function correctly on first try.

Bad trigger: User and Claude had a long conversation but it was just iterating on syntax errors.

Stay alert. When you see the spark of exceptional collaboration, capture it for the team.