limit-thinking

Trace convergence of key variables toward extreme boundaries for scaling decisions.

7|2|Updated Mar 5, 2026
One-click install
npx skills add https://github.com/AndurilCode/craftwork --skill limit-thinking
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: limit-thinking
Source: https://github.com/AndurilCode/craftwork/tree/main/plugins/craftwork-reasoning/skills/limit-thinking
Command: npx skills add https://github.com/AndurilCode/craftwork --skill limit-thinking

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Limit thinking helps you avoid planning based on a snapshot by showing what a system or decision converges to when key variables are pushed toward their extremes.

Core Features & Use Cases

  • Convergence-focused evaluation: Reframes “what is the value now?” into “what does this approach at scale?” to surface asymptotes, phase transitions, reversals, and divergence.
  • Convergence trace workflow: Guides you to incrementally push variables (slightly → midpoint → near-limit → limit) to identify where naive expectations fail.
  • Operating point extraction: Turns limit behavior into a practical recommended operating range with early warning signals for inflection, diminishing returns, or collapse.

Use case example: Before rolling out “automation at 100%” for an agent workflow, trace the key variables (automation level, review throughput, and human attention) to see whether performance converges, hits an asymptotic ceiling, or collapses due to atrophy or feedback-loop failures.

Quick Start

Use limit-thinking when you are deciding whether to scale a plan by asking what the outcomes converge to as adoption, automation, time horizon, or cost are pushed toward an extreme.

Frequently Asked Questions about limit-thinking

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

FAQPage Schema
How do I evaluate a scaling strategy before pushing adoption to the maximum limit?

Evaluate a scaling strategy by identifying 2–3 high-leverage variables and incrementally tracing their convergence toward extreme boundaries to surface asymptotes, phase transitions, or collapse. This reveals the true operational ceiling rather than relying on a snapshot.

What is limit analysis and how does it model system trajectories?

Limit analysis models system trajectories by pushing key variables toward extremes to determine what outcomes converge to, rather than evaluating current snapshots. It surfaces asymptotic ceilings, phase transitions, and divergence in coupled parameters across adoption, automation, time, or cost.

How do I find the optimal operating point for an automation rollout plan?

Find the optimal operating point by extracting the delta between naive expectations and actual limit behavior during a convergence trace, then translating that limit behavior into a recommended operating range with early warning signals for inflection or diminishing returns.

When should I use trajectory modeling instead of snapshot evaluation for growth strategy design?

Use trajectory modeling for growth strategy design when you need to see what a system converges to at scale, rather than its current state. It is essential when pushing adoption, automation, time horizon, or cost toward extremes could trigger phase transitions or feedback-loop failures.

Can limit thinking identify early warning signals before a system hits an asymptotic ceiling?

Yes, limit thinking identifies early warning signals by incrementally tracing variables toward their limits and extracting the deltas between naive and actual limits. This process detects inflection points, diminishing returns, or collapse before the system reaches its asymptotic ceiling.

What are the limitations of using limit analysis for rollout plans?

Limit analysis for rollout plans requires identifying only 2–3 highest-leverage variables, meaning it may overlook secondary coupled parameters. It focuses on convergence behavior and phase transitions, so it may not capture short-term snapshot volatility or variables outside the selected high-leverage set.