One-click install
npx skills add https://github.com/marktoda/cohesive --skill audit-substrate
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: audit-substrate
Source: https://github.com/marktoda/cohesive/tree/main/skills/audit-substrate
Command: npx skills add https://github.com/marktoda/cohesive --skill audit-substrate

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates the risk of unknown missing institutional knowledge in codebases, which causes inconsistent changes, avoidable bugs, and wasted time re-discovering implicit rules during refactors, onboarding, or code reviews.

Core Features & Use Cases

  • Single-pass substrate audit: Scans the entire repo or a named subsystem to inventory missing memory without dispatching extra reviewer agents, saving token costs.
  • Prioritized gap reporting: Ranks missing substrate by leverage, focusing only on gaps that would prevent the most predictable future bugs rather than enumerating every minor missing artifact.
  • Use case: Run this before a major refactor, when onboarding to a new codebase, or during periodic repo health checks to identify what specs, invariants, behavior matrices, gotchas, and enforcement rules are absent.

Quick Start

Use the audit-substrate skill to scan the current repository for all missing substrate and generate a prioritized list of high-leverage gaps to address.

Frequently Asked Questions about audit-substrate

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

FAQPage Schema
How do I audit a codebase for missing institutional knowledge before a refactor?

To audit a codebase for missing institutional knowledge before a refactor, you can run a single-pass substrate audit that scans the repo to inventory absent specs, invariants, behavior matrices, and gotchas, generating a prioritized report of high-leverage gaps to address.

What is a substrate audit and how does it identify codebase memory gaps?

A substrate audit is a repository-wide or subsystem-level scan that identifies missing codebase memory, such as absent enforcement rules and implicit behavior matrices, preventing inconsistent changes by surfacing undocumented institutional knowledge.

How do I find missing documentation and implicit invariants during codebase onboarding?

You can find missing documentation and implicit invariants during codebase onboarding by scanning the repository to inventory missing substrate artifacts, which produces a prioritized list of undocumented gotchas and behavior rules mapped to concrete targets.

Can I run a cohesion review on a specific subsystem instead of the entire repository?

Yes, you can run a cohesion review on a specific named subsystem instead of the entire repository, allowing you to target subsystem-level scans to identify missing enforcement rules and institutional memory without dispatching extra reviewer agents.

What is the best way to prioritize missing specs and gotchas in a repository health check?

The best way to prioritize missing specs and gotchas during a repository health check is to rank missing substrate by leverage, focusing on gaps that prevent the most predictable future bugs rather than enumerating every minor missing artifact.

Why does a single-pass repo audit save token costs compared to dispatching reviewer agents?

A single-pass repo audit saves token costs because it scans the entire repository to inventory missing substrate artifacts directly, eliminating the need to dispatch multiple extra reviewer agents to find missing institutional memory and gaps.