skills.rest
chrisliu298
chrisliu298COMMUNITY
Explore all skills by this author →
Website𝕏@chrisliu298GitHub

prism

Dispatch multiple agents to answer one question from diverse analytical lenses and synthesize perspectives.

1|Updated Mar 16, 2026
Search Tags:#agents#orchestration#synthesis#decision-making#parallax#multilens
One-click install
npx skills add https://github.com/chrisliu298/dotfiles --skill prism-chrisliu298
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: prism
Source: https://github.com/chrisliu298/dotfiles/tree/main/agents/extensions/skills/prism
Command: npx skills add https://github.com/chrisliu298/dotfiles --skill prism-chrisliu298

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prism enables sending the same complete question to multiple independent agents and synthesizing their diverse analyses, reducing blind spots and surface tradeoffs.

Core Features & Use Cases

  • Redundancy over division of labor: every agent answers the full question but with different lenses to surface alignment, contradictions, and risk.
  • Parallax cross-model diversity: cross-model analysis to improve coverage and detect model-specific biases.
  • Use cases include architecture decisions, code reviews, and complex design challenges where high-stakes tradeoffs exist.

Quick Start

Ask Prism to run a multi-lens analysis on a non-trivial decision to synthesize diverse perspectives into a single recommendation.

Frequently Asked Questions about prism

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

FAQPage Schema
What is multi-lens reasoning for complex decision-making?▼

Multi-lens reasoning dispatches multiple independent agents to analyze the same complete question from different perspectives, then synthesizes their convergence, divergence, and risks into a single actionable recommendation.

How do I synthesize diverse agent perspectives for architecture decisions?▼

You synthesize diverse agent perspectives by dispatching parallel agents with different analytical lenses to answer the full architecture question, then merging their alignment, contradictions, and risks into actionable recommendations.

When should I use multi-agent synthesis instead of a single agent?▼

Use multi-agent synthesis for non-trivial decisions, ambiguous tradeoffs, or high-stakes changes where a single perspective might miss critical risks, contradictions, or blind spots.

Does multi-agent orchestration require dividing tasks among agents?▼

No, this multi-agent orchestration uses redundancy rather than division of labor, ensuring every agent answers the full question with a different lens to surface alignment and contradictions.

Can cross-model analysis help detect biases in code reviews?▼

Yes, cross-model analysis improves coverage and detects model-specific biases by running parallax diversity across multiple independent agents during tasks like code reviews.

What are the limitations of using multiple independent agents for tradeoff analysis?▼

This approach requires parallel agent prompts and diverse lenses, making it resource-intensive for trivial decisions where a single perspective is sufficient and faster.

Related Skills

Explore More Software Engineering Skills→
flonatflonat
127

multi-perspective

Spawn parallel agents with distinct disciplinary lenses to synthesize research findings.

Community
Advanced
brolagbrolag
11

parallel-verification

Explore multiple reasoning paths in parallel to verify high-stakes solutions.

Community
Advanced
Oscar-Ivan-SalasOscar-Ivan-Salas

parallel-agents

Coordinate domain-specific agents for multi-perspective software project analyses.

Community
Advanced
luyzkkluyzkk

doubt-driven-development

Review a single decision through a bounded adversarial CLAIM-to-STOP process before committing.

Community
Advanced