skills.rest
sakamoto-family-smile
sakamoto-family-smileCOMMUNITY
Explore all skills by this author →
GitHub

agent-architecture-audit

Audit the 12-layer agent stack for wrapper regressions and memory pollution.

Updated Apr 13, 2026
Search Tags:#architecture#debugging#security#audit#agent#diagnose
One-click install
npx skills add https://github.com/sakamoto-family-smile/agent_monorepo --skill agent-architecture-audit-sakamoto-family-smile
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-architecture-audit
Source: https://github.com/sakamoto-family-smile/agent_monorepo/tree/main/.claude/skills/ecc/agent-architecture-audit
Command: npx skills add https://github.com/sakamoto-family-smile/agent_monorepo --skill agent-architecture-audit-sakamoto-family-smile

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Full-stack diagnostics for agent and LLM applications. This skill audits the 12-layer agent stack to uncover wrapper regression, memory pollution, tool discipline failures, hidden repair loops, and rendering corruption, delivering severity-ranked findings with actionable fixes.

Core Features & Use Cases

  • End-to-end health checks across prompts, memory, tool execution, and rendering
  • Structured audit workflow (scope, evidence collection, failure mapping, and fix strategy)
  • Code-first remediation guidance tailored to developer workflows

Quick Start

Run the audit workflow against your agent repository to generate a findings report.

Frequently Asked Questions about agent-architecture-audit

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

FAQPage Schema
How do I diagnose agent architecture failures across my LLM application stack?▼

Agent architecture audits identify failure origins by mapping evidence across the 12-layer stack, from prompt design to rendering delivery, ensuring structured scope and code-first remediation plans for development, staging, and production environments.

What is the best way to fix hidden repair loops and memory pollution in agent applications?▼

Agent architecture audits identify failure origins by mapping evidence across the 12-layer stack, from prompt design to rendering delivery, ensuring structured scope and code-first remediation plans for development, staging, and production environments.

Does agent architecture auditing work for production environments or only development?▼

Agent architecture audits identify failure origins by mapping evidence across the 12-layer stack, from prompt design to rendering delivery, ensuring structured scope and code-first remediation plans for development, staging, and production environments.

How to audit tool discipline failures and wrapper regressions in an agent repository?▼

Agent architecture audits identify failure origins by mapping evidence across the 12-layer stack, from prompt design to rendering delivery, ensuring structured scope and code-first remediation plans for development, staging, and production environments.

When do I need a full-stack diagnostic audit for my LLM agent?▼

Agent architecture audits identify failure origins by mapping evidence across the 12-layer stack, from prompt design to rendering delivery, ensuring structured scope and code-first remediation plans for development, staging, and production environments.

What are the limitations of diagnosing agent failures without a structured audit workflow?▼

Agent architecture audits identify failure origins by mapping evidence across the 12-layer stack, from prompt design to rendering delivery, ensuring structured scope and code-first remediation plans for development, staging, and production environments.

Related Skills

Explore More Software Engineering Skills→
axelcool1234axelcool1234

host-audit

Audit a host against repository policy and live machine facts.

Community
Advanced
linenoizelinenoize
1

review

Review code changes for defects, security risks, and missing test coverage.

Community
Advanced
alo-expalo-exp
6

ai-llm-safety

Design and implement safety controls for LLM-based agent workflows.

Official
Advanced
JKKN-InstitutionsJKKN-Institutions

advanced-debugging

Diagnose Next.js 15 and Supabase issues across UI, Service, and Database layers.

Official
Intermediate