skills.rest
AnastasiyaW
AnastasiyaWCOMMUNITY
Explore all skills by this author →
GitHub

harness-audit

Audit agent harness maturity across five subsystems and generate a prioritized improvement plan.

140|23|Updated Mar 28, 2026
Search Tags:#workflow#optimization#audit#compliance#agent#harness
One-click install
npx skills add https://github.com/AnastasiyaW/codex-claude-code-config --skill harness-audit-anastasiyaw
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: harness-audit
Source: https://github.com/AnastasiyaW/codex-claude-code-config/tree/main/skills/operational/harness-audit
Command: npx skills add https://github.com/AnastasiyaW/codex-claude-code-config --skill harness-audit-anastasiyaw

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This skill addresses the common issue of agent stagnation by evaluating the maturity of your project's agent harness, identifying the specific subsystem bottleneck preventing progress, and providing a clear path to improvement.

Core Features & Use Cases

  • 5-Subsystem Assessment: Evaluates Instructions, State, Verification, Scope, and Lifecycle to provide a comprehensive health score.
  • Bottleneck Identification: Pinpoints exactly which subsystem is limiting your agent's effectiveness.
  • Actionable Roadmap: Generates a prioritized 3-step improvement plan to maximize impact with minimal effort.
  • Use Case: Use this when your agent is failing despite using high-quality models, or when you are onboarding to a new codebase and need to determine if it is ready for long-running autonomous tasks.

Quick Start

Run the harness audit on the current project directory to identify the primary bottleneck and receive a prioritized improvement plan.

Frequently Asked Questions about harness-audit

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

FAQPage Schema
How do I audit my agent harness to find performance bottlenecks?▼

To audit an agent harness, evaluate the maturity of instructions, state, verification, scope, and lifecycle subsystems to identify the primary bottleneck and generate a prioritized improvement plan.

Why does my agent stagnate even when using high-quality models?▼

Agent stagnation often occurs because of an immature project harness rather than model quality. Auditing the five critical harness subsystems identifies the specific bottleneck preventing progress.

What files do I need to run an agent harness audit?▼

Running an agent harness audit requires access to project-level configuration files, specifically CLAUDE.md, feature_list.json, and init.sh, to evaluate the maturity of the five subsystems.

When should I evaluate my project's agent harness maturity?▼

You should evaluate agent harness maturity when onboarding to a new codebase to determine readiness for long-running autonomous tasks, or when your agent is failing despite using high-quality models.

What is the best way to improve agent effectiveness after identifying a bottleneck?▼

The best way to improve agent effectiveness is to follow a prioritized three-step improvement plan generated after evaluating the five harness subsystems to maximize impact with minimal effort.

Related Skills

Explore More Software Engineering Skills→
enkira-aienkira-ai
2

harness-audit

Audit repository scaffolding files and generate a structured health report.

Official
Advanced
AskTinNguyenAskTinNguyen

workflow-improve-skill

Coordinate parallel audit agents to score SKILL.md against a rubric and generate prioritized improvement plans.

Community
Advanced
doitiandoitian
14

dispatching-parallel-agents

Dispatch dedicated agents to investigate independent failures in parallel.

Community
Intermediate
liauw-medialiauw-media
3

dispatching-parallel-agents

Dispatch parallel agents to execute independent tasks concurrently.

Community
Advanced