Skip to main content

Overview

The info command displays skill metadata and token usage estimates. It provides quick insight into a skill’s structure, content size, and expected token costs for discovery and activation.

Usage

Arguments

Path
Path to the skill directory. Defaults to current directory if not specified.

Options

boolean
default:false
Output as JSON for machine-readable format (pipe-friendly).
string
Alias: -fExtract a single field value from the skill metadata.Available fields:
  • name - Skill name
  • description - Skill description
  • license - License type
  • compatibility - Compatibility information
  • structure - List of subfolders
  • body_lines - Number of lines in skill body
  • tokens - Token estimates (returns JSON object)

Examples

Display skill info (default format)

Output:

Get JSON output

Output:

Extract specific field

Use in scripts

Output Fields

Structure Detection

The command automatically detects skill structure:
  • scripts/ - Contains executable scripts
  • references/ - Contains reference documentation
  • assets/ - Contains images or other assets

Token Estimates

Discovery tokens: Estimated tokens for skill name + description
  • Used during skill discovery phase
  • Affects how many skills can be loaded in agent context
Activation tokens: Estimated tokens for full SKILL.md content
  • Used when skill is activated
  • Includes frontmatter, body, and all content
Token estimates use a simplified algorithm and may differ from actual token counts by specific models. Use them as approximations for budgeting.

JSON Output Format

Exit Codes

  • 0: Info retrieved successfully
  • 1: Error occurred (invalid path, missing SKILL.md, unknown field)

Use Cases

Skill Budget Monitoring

Check if skills fit within token budgets:

CI/CD Integration

Validate skill metadata in automated workflows:

Skill Catalog Generation

Generate skill catalogs from metadata:

Notes

Token estimates are based on approximate character-to-token ratios. For precise token counts, use the actual model’s tokenizer.