agent-customization

Inject project-specific context from AGENTS.md into installed agent templates.

95|4|Updated Feb 27, 2026
One-click install
npx skills add https://github.com/aspenkit/aspens --skill agent-customization-aspenkit
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-customization
Source: https://github.com/aspenkit/aspens/tree/main/.agents/skills/agent-customization
Command: npx skills add https://github.com/aspenkit/aspens --skill agent-customization-aspenkit

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Contextualizing agents with project-specific information to improve behavior and alignment during automation and agent customization.

Core Features & Use Cases

  • Context gathering reads AGENTS.md, full .claude/skills/**, and lists guidelines for reference to tailor agent behavior.
  • Agent discovery traverses .claude/agents/ to detect agent definitions and extract names for targeted customization.
  • Read-only interactions ensure Claude invocations use only allowed tools and do not modify code outside the intended outputs.
  • Output handling writes customized agent templates into .claude/agents/ with minimal risk and clear traceability.

Quick Start

Run the customize command to inject project-specific context into your agents.

Frequently Asked Questions about agent-customization

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

FAQPage Schema
How do I inject project context into Claude agents automatically?

To inject project context into Claude agents automatically, run the customize command. This reads AGENTS.md, .claude/skills, and .claude/guidelines to embed repository-specific information directly into agent templates within .claude/agents.

What is the best way to tailor agent behavior to my repository?

The best way to tailor agent behavior is using a read-only customization flow that gathers context from AGENTS.md and .claude/skills. This safely contextualizes agents without modifying code outside the intended output files in .claude/agents.

How does automatic context embedding work for Claude agent templates?

Automatic context embedding works by traversing .claude/agents to discover existing definitions, extracting agent names, and reading project guidelines. Claude invocations then use only allowed tools to write customized templates with clear traceability.

Do I need AGENTS.md to customize agents for my project?

Yes, you need AGENTS.md to define agent behavior for contextual customization. The customization flow requires AGENTS.md and .claude/skills to read project-specific context and apply it safely to your installed agent templates.

Can I customize agents in .claude/agents without modifying my source code?

Yes, you can customize agents without modifying source code through read-only interactions. Claude invocations use only allowed tools to gather context and write customized agent templates exclusively into .claude/agents with minimal risk.

Why are my customized Claude agents not aligning with my project guidelines?

Customized agents may not align if context gathering cannot read your AGENTS.md, .claude/skills, or .claude/guidelines files. Ensure these files exist and define the correct project context for the automatic injection workflow.