agent-md-refactor
softaworks/agent-toolkit
Refactor bloated agent instruction files into organized, linked documentation following progressive disclosure principles.
What is agent-md-refactor?
This skill refactors monolithic agent instruction files (AGENTS.md, CLAUDE.md, etc.) by identifying contradictions, extracting essentials, categorizing remaining content, and splitting into a minimal root file with organized linked documentation. Use it when your agent config files have grown too large or disorganized.
- Identifies and resolves contradictions in instruction files
- Extracts essential content for the root file (project description, commands, critical overrides)
- Groups remaining instructions into logical categories (TypeScript, testing, code style, git workflow, architecture, etc.)
- Creates a file hierarchy with minimal root and linked topic-specific files
- Flags redundant, vague, or outdated instructions for deletion
- Verifies structure with actionable content and working links
How to install agent-md-refactor
npx skills add https://github.com/softaworks/agent-toolkit --skill agent-md-refactorHow to use agent-md-refactor
- 1.Run the skill and provide your current agent instruction file (AGENTS.md, CLAUDE.md, etc.)
- 2.Phase 1: Review and resolve any contradictions the skill identifies
- 3.Phase 2: Confirm which content is essential for the root file
- 4.Phase 3: Review proposed categories and adjust groupings if needed
- 5.Phase 4: Accept the generated file structure and linked files
- 6.Phase 5: Review flagged instructions for deletion and confirm removals
- 7.Verify the root file is under 50 lines and all links work correctly
Use cases
- Refactor a 500+ line CLAUDE.md into a minimal root with 5-6 linked topic files
- Resolve contradictory style guidelines in an AGENTS.md file before splitting
- Organize scattered instructions in COPILOT.md into a progressive disclosure structure
- Clean up outdated or vague instructions that waste token context
- Create a maintainable agent instruction hierarchy for a team project
- Developers managing agent instruction files for Claude, Cursor, or similar coding agents
- Teams with bloated or contradictory agent configuration files
- Projects needing organized, linked documentation for agent behavior
- Anyone using progressive disclosure to keep essential info at root level
agent-md-refactor FAQ
Only information that applies to every single task: one-sentence project description, non-standard package manager or commands, critical overrides, and universal rules. Language-specific conventions, testing guidelines, code style details, and framework patterns belong in linked files.
Aim for 3-8 files. Common categories include typescript.md, testing.md, code-style.md, git-workflow.md, and architecture.md. Consolidate related topics rather than creating too many granular files.
Remove instructions that are redundant (agent already knows), too vague to be actionable, overly obvious, default behavior, or outdated. Examples: 'Write clean code', 'Use TypeScript' (in a TS project), 'Don't introduce bugs'.
Yes. The skill provides a standard structure, but you can adjust category names, folder location (.claude/ vs docs/agent-instructions/), and groupings based on your project needs.
Verify: root file is under 50 lines with only universal info, all links work, no contradictions remain, every instruction is specific and actionable, and each linked file is self-contained.
Full instructions (SKILL.md)
Source of truth, from softaworks/agent-toolkit.
name: agent-md-refactor description: Refactor bloated AGENTS.md, CLAUDE.md, or similar agent instruction files to follow progressive disclosure principles. Splits monolithic files into organized, linked documentation. license: MIT
Agent MD Refactor
Refactor bloated agent instruction files (AGENTS.md, CLAUDE.md, COPILOT.md, etc.) to follow progressive disclosure principles - keeping essentials at root and organizing the rest into linked, categorized files.
Triggers
Use this skill when:
- "refactor my AGENTS.md" / "refactor my CLAUDE.md"
- "split my agent instructions"
- "organize my CLAUDE.md file"
- "my AGENTS.md is too long"
- "progressive disclosure for my instructions"
- "clean up my agent config"
Quick Reference
| Phase | Action | Output |
|---|---|---|
| 1. Analyze | Find contradictions | List of conflicts to resolve |
| 2. Extract | Identify essentials | Core instructions for root file |
| 3. Categorize | Group remaining instructions | Logical categories |
| 4. Structure | Create file hierarchy | Root + linked files |
| 5. Prune | Flag for deletion | Redundant/vague instructions |
Process
Phase 1: Find Contradictions
Identify any instructions that conflict with each other.
Look for:
- Contradictory style guidelines (e.g., "use semicolons" vs "no semicolons")
- Conflicting workflow instructions
- Incompatible tool preferences
- Mutually exclusive patterns
For each contradiction found:
## Contradiction Found
**Instruction A:** [quote]
**Instruction B:** [quote]
**Question:** Which should take precedence, or should both be conditional?
Ask the user to resolve before proceeding.
Phase 2: Identify the Essentials
Extract ONLY what belongs in the root agent file. The root should be minimal - information that applies to every single task.
Essential content (keep in root):
| Category | Example |
|---|---|
| Project description | One sentence: "A React dashboard for analytics" |
| Package manager | Only if not npm (e.g., "Uses pnpm") |
| Non-standard commands | Custom build/test/typecheck commands |
| Critical overrides | Things that MUST override defaults |
| Universal rules | Applies to 100% of tasks |
NOT essential (move to linked files):
- Language-specific conventions
- Testing guidelines
- Code style details
- Framework patterns
- Documentation standards
- Git workflow details
Phase 3: Group the Rest
Organize remaining instructions into logical categories.
Common categories:
| Category | Contents |
|---|---|
typescript.md | TS conventions, type patterns, strict mode rules |
testing.md | Test frameworks, coverage, mocking patterns |
code-style.md | Formatting, naming, comments, structure |
git-workflow.md | Commits, branches, PRs, reviews |
architecture.md | Patterns, folder structure, dependencies |
api-design.md | REST/GraphQL conventions, error handling |
security.md | Auth patterns, input validation, secrets |
performance.md | Optimization rules, caching, lazy loading |
Grouping rules:
- Each file should be self-contained for its topic
- Aim for 3-8 files (not too granular, not too broad)
- Name files clearly:
{topic}.md - Include only actionable instructions
Phase 4: Create the File Structure
Output structure:
project-root/
├── CLAUDE.md (or AGENTS.md) # Minimal root with links
└── .claude/ # Or docs/agent-instructions/
├── typescript.md
├── testing.md
├── code-style.md
├── git-workflow.md
└── architecture.md
Root file template:
# Project Name
One-sentence description of the project.
## Quick Reference
- **Package Manager:** pnpm
- **Build:** `pnpm build`
- **Test:** `pnpm test`
- **Typecheck:** `pnpm typecheck`
## Detailed Instructions
For specific guidelines, see:
- [TypeScript Conventions](.claude/typescript.md)
- [Testing Guidelines](.claude/testing.md)
- [Code Style](.claude/code-style.md)
- [Git Workflow](.claude/git-workflow.md)
- [Architecture Patterns](.claude/architecture.md)
Each linked file template:
# {Topic} Guidelines
## Overview
Brief context for when these guidelines apply.
## Rules
### Rule Category 1
- Specific, actionable instruction
- Another specific instruction
### Rule Category 2
- Specific, actionable instruction
## Examples
### Good
\`\`\`typescript
// Example of correct pattern
\`\`\`
### Avoid
\`\`\`typescript
// Example of what not to do
\`\`\`
Phase 5: Flag for Deletion
Identify instructions that should be removed entirely.
Delete if:
| Criterion | Example | Why Delete |
|---|---|---|
| Redundant | "Use TypeScript" (in a .ts project) | Agent already knows |
| Too vague | "Write clean code" | Not actionable |
| Overly obvious | "Don't introduce bugs" | Wastes context |
| Default behavior | "Use descriptive variable names" | Standard practice |
| Outdated | References deprecated APIs | No longer applies |
Output format:
## Flagged for Deletion
| Instruction | Reason |
|-------------|--------|
| "Write clean, maintainable code" | Too vague to be actionable |
| "Use TypeScript" | Redundant - project is already TS |
| "Don't commit secrets" | Agent already knows this |
| "Follow best practices" | Meaningless without specifics |
Execution Checklist
[ ] Phase 1: All contradictions identified and resolved
[ ] Phase 2: Root file contains ONLY essentials
[ ] Phase 3: All remaining instructions categorized
[ ] Phase 4: File structure created with proper links
[ ] Phase 5: Redundant/vague instructions removed
[ ] Verify: Each linked file is self-contained
[ ] Verify: Root file is under 50 lines
[ ] Verify: All links work correctly
Anti-Patterns
| Avoid | Why | Instead |
|---|---|---|
| Keeping everything in root | Bloated, hard to maintain | Split into linked files |
| Too many categories | Fragmentation | Consolidate related topics |
| Vague instructions | Wastes tokens, no value | Be specific or delete |
| Duplicating defaults | Agent already knows | Only override when needed |
| Deep nesting | Hard to navigate | Flat structure with links |
Examples
Before (Bloated Root)
# CLAUDE.md
This is a React project.
## Code Style
- Use 2 spaces
- Use semicolons
- Prefer const over let
- Use arrow functions
... (200 more lines)
## Testing
- Use Jest
- Coverage > 80%
... (100 more lines)
## TypeScript
- Enable strict mode
... (150 more lines)
After (Progressive Disclosure)
# CLAUDE.md
React dashboard for real-time analytics visualization.
## Commands
- `pnpm dev` - Start development server
- `pnpm test` - Run tests with coverage
- `pnpm build` - Production build
## Guidelines
- [Code Style](.claude/code-style.md)
- [Testing](.claude/testing.md)
- [TypeScript](.claude/typescript.md)
Verification
After refactoring, verify:
- Root file is minimal - Under 50 lines, only universal info
- Links work - All referenced files exist
- No contradictions - Instructions are consistent
- Actionable content - Every instruction is specific
- Complete coverage - No instructions were lost (unless flagged for deletion)
- Self-contained files - Each linked file stands alone
Related skills
More from softaworks/agent-toolkit and the wider catalog.
mermaid-diagrams
Create professional software diagrams from text using Mermaid syntax.
humanizer
Remove AI-generated writing patterns to make text sound more natural and human.
qa-test-planner
Generate test plans, test cases, regression suites, and bug reports with Figma design validation for QA engineers.
writing-clearly-and-concisely
Apply Strunk's timeless rules to write clearer, stronger prose for humans.
database-schema-designer
Design production-ready SQL and NoSQL database schemas with normalization, indexing, and migration strategies.
skill-judge
Evaluate Agent Skill design quality against specifications and best practices with multi-dimensional scoring.