Monitors Claude's responses for struggle signals and suggests escalation to deep-thinking agents when complexity exceeds comfortable reasoning capacity.
Monitors your responses for struggle signals like hedging or contradictions, then suggests escalating to deep-thinking agents when handling architectural decisions, ambiguous requirements, or multi-domain problems that exceed comfortable reasoning capacity.
/plugin marketplace add ANcpLua/ancplua-claude-plugins/plugin install metacognitive-guard@ancplua-claude-pluginsThis skill inherits all available tools. When active, it can use any tool Claude has access to.
This skill provides awareness of the struggle detection system and guidance on when to proactively engage deep-thinking resources.
Even before the struggle detector triggers, consider spawning deep-think-partner when:
Architectural decisions with competing constraints
Ambiguous requirements requiring interpretation
Multi-domain synthesis required
Edge case analysis needed
Before responding to complex questions, ask yourself:
If you answer "no" to any of these, consider proactive escalation.
Use the Task tool with the deep-think-partner agent:
Task tool:
subagent_type: deep-think-partner
prompt: [Detailed problem statement with all constraints]
description: [3-5 word summary]
Include:
User asks: "Should we use Redis or PostgreSQL for session storage?"
Self-assessment: Multiple valid approaches, depends on constraints not yet explored, "it depends" isn't helpful.
Escalation:
Task tool:
subagent_type: deep-think-partner
prompt: |
Context: Web application with 10k concurrent users, existing PostgreSQL database.
Question: Redis vs PostgreSQL for session storage.
Constraints: Team has PostgreSQL expertise, no Redis experience.
Must handle session expiry. Cost-sensitive.
Success: Clear recommendation with migration path.
description: Analyze session storage options
The automatic detector looks for these patterns in your responses:
| Signal | What It Means | Better Approach |
|---|---|---|
| Hedging | Uncertainty about recommendation | Escalate for deeper analysis |
| Deflecting | Avoiding commitment with questions | Answer then ask clarifying questions |
| Verbose | Rambling without concrete output | Structure response, include code/tables |
| Contradiction | Changed position mid-response | Stop, think, give one coherent answer |
| Apologetic | Previous response was wrong | Acknowledge, correct, move forward |
| Weaseling | Non-committal to avoid being wrong | Make a recommendation with confidence level |
When deep-think-partner returns its analysis:
Track your struggle detection rate to improve:
Self-awareness of your own patterns helps calibrate both the detector and your escalation instincts.
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.