Negentropy Lens

A decision-support framework that evaluates systems and architectures by asking whether they're decaying into complexity debt or compounding value over time. You'd use this when reviewing technical designs, choosing between options, or auditing existing systems. It walks through five phases: mapping the system boundary, diagnosing entropy vs negentropy indicators, surfacing tacit knowledge gaps (the unwritten assumptions and workarounds), evaluating compounding potential, and then challenging the framing. The tacit knowledge probe is non-negotiable, which is the right call. Most architecture decisions rest on unstated assumptions about how things actually work versus how documentation claims they work.

skills.sh installs: 3.4K
GitHub stars: 491
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add bencium/bencium-marketplace --skill negentropy-lens --agent claude-code

Installs into .claude/skills of the current project.

Files

Negentropy Lens

A thinking framework for evaluating decisions, systems, and architectures through two fundamental system states: entropy (decay, disorder, complexity debt) and negentropy (growth, compounding value, increasing order).

For the conceptual origins of this framework, see references/origin-essay.md.

Core Principle

Every system exists in one of two states. Every decision either accelerates entropy or drives negentropy. There is no neutral. Inaction is entropic. The goal is not to eliminate entropy — it is to recognize which state a system is in, surface what is hidden, and make deliberate choices about direction.

Term Definitions

On first use in every output, define these three terms inline using parentheses:

Entropy (the natural tendency of systems toward decay, disorder, and complexity without value)
Negentropy (the deliberate reversal of decay — growth, compounding value, increasing order)
Tacit knowledge (the unwritten, unspoken knowledge of how things actually work — assumptions, workarounds, and institutional memory that never make it into documentation)

SKILL.md on GitHub