Skip to content
Back to skills

2626 Skills Reference Manual 654435f8

ASecurity

**Version**: 2.4.0 **Last Updated**: 2025-12-19 **Author**: Intent Solutions (Jeremy Longshore) ---

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
devopsjavascripttypescriptpythonrustgojavashellbashsqldocker

Works with

  • claude code
  • terminal
  • cli
  • api

Security analysis

A92/100
  • mediumInstalls packages at runtime which could introduce malicious dependencies

Pro shows the line behind each finding and how to fix it

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 2626-skills-reference-manual_654435f8 --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 2626 Skills Reference Manual 654435f8?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for 2626 Skills Reference Manual 654435f8
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-2626-skills-reference-manual-654435f8/badge)](https://www.skillsdirectory.com/skills/tools-only-2626-skills-reference-manual-654435f8)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

SKILL.md
# Claude Code Skills Reference Manual

**Version**: 2.4.0
**Last Updated**: 2025-12-19
**Author**: Intent Solutions (Jeremy Longshore)

---

## Table of Contents

1. [YAML Frontmatter Schema](#1-yaml-frontmatter-schema)
2. [Description Field Specification](#2-description-field-specification)
3. [Allowed Tools Reference](#3-allowed-tools-reference)
4. [Naming Conventions](#4-naming-conventions)
5. [Instruction Body Structure](#5-instruction-body-structure)
6. [Progressive Disclosure Patterns](#6-progressive-disclosure-patterns)
7. [Validation Rules](#7-validation-rules)
8. [Enterprise Extension Fields](#8-enterprise-extension-fields)
9. [Complete Templates](#9-complete-templates)
10. [Anti-Patterns to Avoid](#10-anti-patterns-to-avoid)

---

## 1. YAML Frontmatter Schema

### Anthropic Required Fields

```yaml
---
name: skill-name-kebab-case          # REQUIRED - max 64 chars
description: |                        # REQUIRED - max 1024 chars
  What this skill does. When to use it.
  Trigger phrases for discovery.
---
```

### Enterprise Required Fields

```yaml
---
name: skill-name-kebab-case
description: |
  Primary capability. Secondary features.
  Use when [scenarios]. Trigger with "[phrases]".
allowed-tools:                        # REQUIRED (Enterprise)
  - Read
  - Write
  - Bash
version: 1.0.0                        # REQUIRED (Enterprise) - semver
author: Name <email@domain.com>       # REQUIRED (Enterprise)
license: MIT                          # REQUIRED (Enterprise)
tags:                                 # RECOMMENDED
  - category
  - domain
---
```

### Complete Schema Reference

| Field | Type | Required | Max Length | Validation |
|-------|------|----------|------------|------------|
| `name` | string | Yes (Anthropic) | 64 chars | kebab-case, no XML, no reserved words |
| `description` | string | Yes (Anthropic) | 1024 chars | Non-empty, no XML tags |
| `allowed-tools` | array/CSV | Yes (Enterprise) | - | Valid tool names |
| `version` | string | Yes (Enterprise) | - | Semver format (X.Y.Z) |
| `author` | string | Yes (Enterprise) | - | Name <email> format |
| `license` | string | Yes (Enterprise) | - | Valid license (MIT, Apache-2.0, etc.) |
| `tags` | array | Recommended | - | Lowercase strings |

### Reserved Words (Cannot Use in Name)

- `anthropic`
- `claude`

---

## 2. Description Field Specification

### The Description Formula

```
[Action Verbs] + [Specific Capabilities] + [Use When] + [Trigger Phrases]
```

### Action Verbs Reference

| Category | Action Verbs |
|----------|-------------|
| **Data Operations** | Extract, analyze, parse, transform, convert, merge, split, validate, filter, aggregate |
| **Creation** | Generate, create, build, produce, synthesize, compose, scaffold, initialize |
| **Modification** | Edit, update, refactor, optimize, fix, enhance, migrate, patch, upgrade |
| **Analysis** | Review, audit, scan, inspect, diagnose, profile, assess, evaluate, benchmark |
| **Operations** | Deploy, execute, run, configure, install, setup, provision, orchestrate |
| **Documentation** | Document, explain, summarize, annotate, describe, catalog, index |
| **Testing** | Test, verify, validate, check, assert, measure, monitor |
| **Security** | Secure, encrypt, authenticate, authorize, protect, harden, sanitize |

### Description Patterns

**Pattern 1: Action-Focused (RECOMMENDED)**
```yaml
description: |
  Extract text and tables from PDFs, fill forms, merge documents.
  Use when working with PDF files or when user mentions PDFs, forms, or document extraction.
```

**Pattern 2: Capability + Trigger**
```yaml
description: |
  Kubernetes pod debugging and troubleshooting toolkit.
  Use when pods crash, fail to start, or exhibit unexpected behavior.
  Trigger with "debug pod", "pod failing", "container crash".
```

**Pattern 3: Domain-Specific**
```yaml
description: |
  Analyze SQL query performance and suggest optimizations.
  Use for slow queries, missing indexes, N+1 problems, and query plan analysis.
  Trigger with "optimize query", "slow SQL", "query performance".
```

**Pattern 4: Multi-Capability**
```yaml
description: |
  Comprehensive Docker Compose management: generate configs, validate syntax, optimize services.
  Use when creating docker-compose.yaml, debugging container issues, or optimizing multi-service deployments.
  Trigger with "docker compose", "container orchestration", "multi-container".
```

### Good vs Bad Examples

| Quality | Example | Problem |
|---------|---------|---------|
| BAD | `Helps with documents` | Too vague, no triggers |
| BAD | `Processes data` | No specificity |
| BAD | `I can help you with PDFs` | First person |
| BAD | `You can use this for files` | Second person |
| GOOD | `Extract text and tables from PDF files, fill forms, merge documents. Use when working with PDF files or when the user mentions PDFs, forms, or document extraction.` | Specific, third person, has triggers |

### Critical Rules

1. **Always third person** - Never "I can help" or "You can use"
2. **Include file types** - .pdf, .xlsx, .yaml, .json, .ts
3. **Include domain keywords** - Kubernetes, SQL, Docker, Git
4. **Define boundaries** - What it cannot do
5. **Max 1024 characters**
6. **Include trigger phrases** - What users might say

---

## 3. Allowed Tools Reference

### Core Tools

| Tool | Purpose | Example |
|------|---------|---------|
| `Read` | Read file contents | Read configuration files |
| `Write` | Create/overwrite files | Generate new files |
| `Edit` | Modify existing files | Update code |
| `Grep` | Search file contents | Find patterns |
| `Glob` | Find files by pattern | Locate files |
| `Bash` | Execute shell commands | Run scripts |
| `WebFetch` | Fetch web content | API calls |
| `WebSearch` | Search the web | Research |

### Tool Permission Categories

**Read-Only Analysis**
```yaml
allowed-tools:
  - Read
  - Grep
  - Glob
```

**Code Editing**
```yaml
allowed-tools:
  - Read
  - Write
  - Edit
  - Grep
  - Glob
  - Bash
```

**Web Research**
```yaml
allowed-tools:
  - Read
  - WebFetch
  - WebSearch
  - Grep
```

**Database Operations**
```yaml
allowed-tools:
  - Read
  - Write
  - Bash
  - Grep
```

**Testing**
```yaml
allowed-tools:
  - Read
  - Bash
  - Grep
  - Glob
```

### Bash Tool Scoping

```yaml
# Full Bash access
allowed-tools:
  - Bash

# Scoped Bash (specific command categories)
allowed-tools:
  - Bash(git:*)      # Git commands only
  - Bash(npm:*)      # NPM commands only
  - Bash(docker:*)   # Docker commands only
```

### Tool Format Options

**Array Format (RECOMMENDED)**
```yaml
allowed-tools:
  - Read
  - Write
  - Bash
```

**CSV Format**
```yaml
allowed-tools: Read, Write, Bash
```

---

## 4. Naming Conventions

### Name Field Rules

| Rule | Requirement |
|------|-------------|
| Characters | Lowercase letters, numbers, hyphens only |
| Format | `kebab-case` |
| Max Length | 64 characters |
| Restrictions | No XML tags, no reserved words |

### Naming Patterns

**Gerund Form (RECOMMENDED)**
```
processing-pdfs
analyzing-spreadsheets
managing-databases
testing-code
```

**Noun Phrases**
```
pdf-processor
spreadsheet-analyzer
database-manager
code-tester
```

**Action-Oriented**
```
process-pdfs
analyze-spreadsheets
manage-database
test-code
```

### Bad Names to Avoid

| Name | Problem |
|------|---------|
| `helper` | Too vague |
| `utils` | Not descriptive |
| `tools` | Generic |
| `my-awesome-skill` | Not professional |
| `anthropic-helper` | Reserved word |
| `claude-tools` | Reserved word |
| `PDF_Processor` | Wrong case |
| `pdf processor` | Spaces not allowed |

---

## 5. Instruction Body Structure

### Standard Structure

```markdown
---
name: skill-name
description: |
  Description here.
allowed-tools: "Read, Write, Bash(git:*)"
version: 1.0.0
author: Name <email>
license: MIT
---

# Skill Name

Brief purpose statement (1-2 sentences).

## Overview

What this skill does, when to use it, key capabilities.
- Capability 1
- Capability 2
- Capability 3

## Prerequisites

- Required tool: `tool-name`
- Environment variable: `ENV_VAR`
- Dependency: `package-name`

## Instructions

### Step 1: Verb Action

Clear, imperative instructions.

```bash
command example
```

### Step 2: Verb Action

More instructions with code examples.

### Step 3: Verb Action

Final steps.

## Output

What artifacts this skill produces:
- File: `output.json`
- Report: Summary in terminal
- Modified: Updated configuration

## Error Handling

### Common Error 1

**Cause**: Description of cause
**Solution**: How to fix

### Common Error 2

**Cause**: Description
**Solution**: Fix steps

## Examples

### Example 1: Basic Usage

**Input**:
```
User request example
```

**Output**:
```
Expected result
```

### Example 2: Advanced Usage

More complex scenario.

## Resources

- [Official Docs](https://example.com)
- See `{baseDir}/reference.md` for details
```

### Body Length Guidelines

| Content | Max Lines | Notes |
|---------|-----------|-------|
| SKILL.md body | 500 lines | Split if longer |
| Quick start | 20-30 lines | Essential info only |
| Full reference | Separate file | Use progressive disclosure |

---

## 6. Progressive Disclosure Patterns

### Pattern 1: High-Level Guide with References

```markdown
# PDF Processing

## Quick Start

Extract text with pdfplumber:
```python
import pdfplumber
with pdfplumber.open("file.pdf") as pdf:
    text = pdf.pages[0].extract_text()
```

## Advanced Features

**Form filling**: See [FORMS.md](FORMS.md) for complete guide
**API reference**: See [REFERENCE.md](REFERENCE.md) for all methods
**Examples**: See [EXAMPLES.md](EXAMPLES.md) for common patterns
```

### Pattern 2: Domain-Specific Organization

```
skill-directory/
├── SKILL.md (overview and navigation)
└── reference/
    ├── finance.md (revenue metrics)
    ├── sales.md (pipeline data)
    ├── product.md (usage analytics)
    └── marketing.md (campaigns)
```

### Pattern 3: Conditional Details

```markdown
# Document Processing

## Creating Documents

Use docx-js for new documents. See [DOCX-JS.md](DOCX-JS.md).

## Editing Documents

For simple edits, modify the XML directly.

**For tracked changes**: See [REDLINING.md](REDLINING.md)
**For OOXML details**: See [OOXML.md](OOXML.md)
```

### Key Rules

1. **Keep references one level deep** - All reference files link from SKILL.md
2. **Add table of contents** - For files > 100 lines
3. **Use `{baseDir}`** - For portable file references

---

## 7. Validation Rules

### Field Validation

| Field | Validation |
|-------|------------|
| `name` | Non-empty, max 64 chars, kebab-case, no reserved words |
| `description` | Non-empty, max 1024 chars, no XML tags |
| `allowed-tools` | Valid tool names from known list |
| `version` | Semver format (X.Y.Z) |
| `author` | Non-empty, ideally `Name <email>` format |
| `license` | Valid SPDX license identifier |

### Content Validation

| Check | Rule |
|-------|------|
| Hardcoded paths | No `/tmp/`, `/home/user/`, absolute paths |
| Description quality | Must include action verbs |
| File references | Use `{baseDir}` for bundled files |
| Tool references | Must be in `allowed-tools` |

### Validator Script

```bash
# Validate all skills
python3 scripts/validate-skills-schema.py

# Validate specific directory
python3 scripts/validate-skills-schema.py plugins/devops/

# Fix enterprise fields
python3 scripts/fix-skills-enterprise.py
```

---

## 8. Enterprise Extension Fields

### Intent Solutions Standard

```yaml
version: 1.0.0                                    # Required
author: Jeremy Longshore <jeremy@intentsolutions.io>  # Required
license: MIT                                      # Required
tags:                                             # Recommended
  - category
  - domain
```

### Tag Taxonomy

**Primary Categories**
- `devops`, `security`, `testing`, `performance`
- `database`, `api`, `documentation`, `ai-ml`

**Domains**
- `kubernetes`, `docker`, `terraform`, `aws`, `gcp`, `azure`
- `python`, `typescript`, `javascript`, `rust`, `go`
- `postgresql`, `mongodb`, `redis`, `elasticsearch`

**Actions**
- `debugging`, `optimization`, `generation`, `analysis`
- `migration`, `deployment`, `monitoring`, `validation`

---

## 9. Complete Templates

### Basic Skill Template

```yaml
---
name: basic-skill-name
description: |
  Primary capability as action verb. Secondary features.
  Use when [2-3 specific scenarios].
  Trigger with "phrase 1", "phrase 2", "phrase 3".
allowed-tools:
  - Read
  - Grep
  - Glob
version: 1.0.0
author: Jeremy Longshore <jeremy@intentsolutions.io>
license: MIT
tags:
  - category
  - domain
---

# Basic Skill Name

Brief purpose statement.

## Overview

What this skill does and when to use it.

## Instructions

### Step 1: Analyze

Read and understand the input.

### Step 2: Process

Apply the skill logic.

### Step 3: Output

Return results to user.

## Examples

### Example 1: Basic Usage

Input and expected output.
```

### Advanced Skill Template (with scripts)

```yaml
---
name: advanced-skill-name
description: |
  Comprehensive toolkit for [domain]. Analyze, transform, and generate.
  Use when working with [file types] or when user needs [capabilities].
  Trigger with "analyze [domain]", "generate [output]", "optimize [target]".
allowed-tools:
  - Read
  - Write
  - Edit
  - Bash
  - Grep
  - Glob
version: 1.0.0
author: Jeremy Longshore <jeremy@intentsolutions.io>
license: MIT
tags:
  - devops
  - automation
---

# Advanced Skill Name

Comprehensive description of purpose.

## Overview

- Capability 1: Description
- Capability 2: Description
- Capability 3: Description

## Prerequisites

- Tool: `required-cli-tool`
- Package: `required-package`

## Workflow

### Phase 1: Analysis

1. Run analysis script:
```bash
python {baseDir}/scripts/analyze.py input.file
```

2. Review output in `analysis.json`

### Phase 2: Transformation

1. Apply transformations:
```bash
python {baseDir}/scripts/transform.py analysis.json
```

### Phase 3: Validation

1. Validate results:
```bash
python {baseDir}/scripts/validate.py output.file
```

## Scripts Reference

| Script | Purpose |
|--------|---------|
| `analyze.py` | Initial analysis |
| `transform.py` | Apply changes |
| `validate.py` | Verify output |

## Error Handling

### Error: Invalid Input

**Cause**: Input file malformed
**Solution**: Validate input format first

## Examples

### Example 1: Full Workflow

Complete example with input, commands, and output.
```

### Read-Only Analysis Template

```yaml
---
name: analysis-skill-name
description: |
  Analyze and report on [domain]. Scan, inspect, and assess.
  Use when reviewing [targets] or auditing [systems].
  Trigger with "analyze [domain]", "audit [target]", "review [subject]".
allowed-tools:
  - Read
  - Grep
  - Glob
version: 1.0.0
author: Jeremy Longshore <jeremy@intentsolutions.io>
license: MIT
tags:
  - analysis
  - audit
---

# Analysis Skill Name

Read-only analysis and reporting.

## Overview

This skill analyzes without modifying:
- Scans for patterns
- Identifies issues
- Generates reports

## Analysis Steps

### Step 1: Gather Data

Read relevant files and configurations.

### Step 2: Analyze Patterns

Search for specific patterns and issues.

### Step 3: Generate Report

Produce summary with findings.

## Output Format

```
## Analysis Report

### Summary
- Total files scanned: X
- Issues found: Y
- Recommendations: Z

### Details
...
```
```

---

## 10. Anti-Patterns to Avoid

### Description Anti-Patterns

| Anti-Pattern | Example | Fix |
|--------------|---------|-----|
| First person | "I can help you with PDFs" | "Extract text from PDF files" |
| Second person | "You can use this for..." | "Use when working with..." |
| Too vague | "Helps with documents" | "Extract, merge, and fill PDF forms" |
| No triggers | "PDF processing tool" | Add "Use when... Trigger with..." |
| Too long | >1024 characters | Condense to essentials |

### Structure Anti-Patterns

| Anti-Pattern | Problem | Fix |
|--------------|---------|-----|
| Deep nesting | References 3+ levels deep | Keep 1 level deep |
| No ToC | Long files hard to navigate | Add table of contents |
| Windows paths | `scripts\helper.py` | Use forward slashes |
| Absolute paths | `/home/user/file.txt` | Use `{baseDir}` |
| Magic numbers | `TIMEOUT = 47` | Document why this value |

### Tool Anti-Patterns

| Anti-Pattern | Problem | Fix |
|--------------|---------|-----|
| `Bash(*)` | Invalid wildcard | Use `Bash` or `Bash(cmd:*)` |
| Overly permissive | All tools for read-only skill | Limit to needed tools |
| Missing tools | Uses Bash but not listed | Add to allowed-tools |

### Content Anti-Patterns

| Anti-Pattern | Problem | Fix |
|--------------|---------|-----|
| Assumes installation | "Use the pdf library" | "Install with `pip install pypdf`" |
| Time-sensitive | "Before August 2025, use old API" | Use "old patterns" section |
| Inconsistent terms | Mix "field", "box", "element" | Pick one, use consistently |
| Too many options | "Use pypdf, or pdfplumber, or..." | Provide default, mention alternatives |

---

## Appendix A: Quick Reference Card

```yaml
# SKILL.md Quick Reference

---
name: kebab-case-name                 # Max 64 chars, required
description: |                        # Max 1024 chars, required
  [Action verb] [capability]. [More features].
  Use when [scenarios].
  Trigger with "[phrase1]", "[phrase2]".
allowed-tools:                        # Enterprise required
  - Read
  - Write
  - Bash
version: 1.0.0                        # Enterprise required
author: Name <email>                  # Enterprise required
license: MIT                          # Enterprise required
tags: [category, domain]              # Recommended
---

# Skill Name

## Overview
What it does.

## Instructions
### Step 1: Verb
Instructions.

## Examples
### Basic Usage
Example.
```

---

## Appendix B: Validation Checklist

- [ ] Name is kebab-case, max 64 chars
- [ ] Description is max 1024 chars
- [ ] Description uses action verbs
- [ ] Description includes trigger phrases
- [ ] Description is third person
- [ ] allowed-tools lists only needed tools
- [ ] version follows semver (X.Y.Z)
- [ ] author is `Name <email>` format
- [ ] license is valid SPDX identifier
- [ ] No hardcoded paths
- [ ] File references use `{baseDir}`
- [ ] Body under 500 lines
- [ ] Complex content in separate files

---

**Document Version**: 2.4.0
**Maintained By**: Intent Solutions
**Source**: Anthropic Official Documentation + Enterprise Extensions

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.