skills.rest
Wondermonger-daydreaming
Wondermonger-daydreamingCOMMUNITY
Explore all skills by this author →
𝕏@fireandvisionGitHub

divert

Surface tail outputs and alternative reasoning paths in Claude.

6|1|Updated May 30, 2026
Search Tags:#LLM#cross-domain#creative-thinking#tail-output#divert#prompt-diversion#non-modal
One-click install
npx skills add https://github.com/Wondermonger-daydreaming/claude-skills-library --skill divert
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: divert
Source: https://github.com/Wondermonger-daydreaming/claude-skills-library/tree/main/skills/divert
Command: npx skills add https://github.com/Wondermonger-daydreaming/claude-skills-library --skill divert

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Recoding-Decoding exposes hidden knowledge in language models by allowing users to surface tail outputs and alternative perspectives that standard decoding hides.

Core Features & Use Cases

  • Tail-access generation: surface non-modal outputs by priming prompts and diversification stems.
  • Mode switches: supports Default (--random), Thick (--thick [concept]), Compare (--compare), Blind (--blind), Chain (--chain [n]), Collision (--collision [A] × [B]).
  • Modifier integration: can be used as a prefix modifier to other skills like /poetry or /paper-riffing.

Quick Start

Invoke the divert mode with a prompt like '/divert' to surface alternative outputs.

Frequently Asked Questions about divert

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I surface hidden reasoning paths and tail outputs in Claude?▼

To surface hidden reasoning paths and tail outputs in Claude, you use prompt diversion techniques to access non-modal insights during brainstorming and cross-domain analysis. This bypasses standard decoding to reveal alternative generation perspectives.

What is the best way to trigger alternative outputs for cross-domain analysis?▼

Triggering alternative outputs for cross-domain analysis requires applying specific diversification modes like --random, --thick, or --collision. These switches prime prompts with alternative stems to generate non-modal perspectives.

Can I use prompt diversion as a modifier with other creative thinking skills?▼

Yes, prompt diversion can be integrated as a prefix modifier to other creative thinking skills. Applying it before another skill alters the base generation, surfacing tail outputs within the downstream workflow.

When do I need to use the collision or chain modes for creative ideation?▼

You need collision or chain modes for creative ideation when exploring complex cross-domain intersections. Collision forces synthesis between two concepts, while chain links sequential alternative reasoning paths to deepen non-modal insights.

Does standard LLM decoding hide alternative perspectives during brainstorming?▼

Yes, standard LLM decoding hides alternative perspectives during brainstorming by favoring high-probability modal outputs. Diverting the output exposes tail generation possibilities, revealing hidden knowledge and non-modal reasoning paths.

Related Skills

Explore More Design & Creative Skills→
frankxaifrankxai
6

arcanea-story-weave

Craft narratives using scene architecture and pacing-control techniques for novels and screenplays.

Community
Basic
d-o-hubd-o-hub
11

analysis-swarm

Orchestrate three AI personas for multi-perspective code analysis.

Community
Advanced
sanjay3290sanjay3290
392

imagen

Generate images from text prompts using Google Gemini.

Community
Intermediate
GPTtangGPTtang
4

prompt-engineering-patterns

Implement few-shot learning and chain-of-thought patterns for structured LLM outputs.

Community
Advanced