How does this skill prevent AI hallucinations?

It enforces a 'slow is fast' approach, requiring Claude to perform line-by-line analysis, cite specific evidence, and explicitly state when information is missing or unclear rather than guessing.

Does this skill automatically find vulnerabilities?

No, this skill is focused exclusively on the context-building phase. It helps Claude understand the code deeply so that subsequent vulnerability-hunting phases are more accurate and less prone to hallucinations.

What analysis methodologies does it use?

The skill applies First Principles reasoning along with the '5 Whys' and '5 Hows' frameworks to interrogate every function's purpose, inputs, and side effects.

When is the best time to activate this skill?

It should be used during the initial orientation and deep-dive phases of a project, specifically before you begin looking for bugs, writing fixes, or modeling exploits.

Can it handle external contract or library calls?

Yes. It treats internal calls as a continuous execution flow and analyzes external 'black-box' calls using an adversarial model to identify potential risks like reentrancy or unexpected return values.

Audit Context Builder

Name: Audit Context Builder
Author: monmacllcapp

bymonmacllcapp

0•

보안 및 테스팅

Enables ultra-granular, line-by-line code analysis to build deep architectural context before vulnerability or bug finding.

The Audit Context Builder is a specialized skill designed to transform Claude's approach to complex codebases by enforcing a rigorous, bottom-up analysis phase. Before jumping to conclusions or hunting for bugs, this skill guides Claude through a structured micro-analysis process—utilizing First Principles, the '5 Whys', and the '5 Hows'—to map out invariants, state transitions, and cross-functional dependencies. By establishing a stable, evidence-based mental model and documenting implicit assumptions, it significantly reduces AI hallucinations and ensures that subsequent security reviews or architectural audits are grounded in a comprehensive understanding of the system's logic.

주요 기능

01Adversarial modeling of external calls and black-box interfaces

02Application of First Principles and '5 Whys' reasoning at a micro scale

03Ultra-granular line-by-line and block-by-block code analysis

04Automated mapping of system-wide invariants and state transitions

050 GitHub stars

06Structured mental model persistence to prevent context degradation

사용 사례

01Verification of invariants and assumptions before performing major refactors

02Deep architectural mapping before beginning a security audit or threat model

03Systematic comprehension of legacy codebases with complex state machines

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add monmacllcapp/skill-forks audit-context-building

For use in Claude.ai and ChatGPT

Download Skill