Skip to content
This repository was archived by the owner on Jul 17, 2026. It is now read-only.

Latest commit

 

History

History
184 lines (129 loc) · 5.55 KB

ai-agents.md

File metadata and controls

184 lines (129 loc) · 5.55 KB

Using LifterLMS CLI with AI Agents

This guide covers how to use the LifterLMS CLI effectively with AI coding assistants like Claude Code, Cursor, Codex, and similar tools.

Why CLI for AI Agents?

AI agents that have shell access (Claude Code, Codex, Cursor terminal) can use the LifterLMS CLI directly — no MCP server or API configuration needed. The CLI runs inside WordPress, so it has full access to all LifterLMS data with proper permission handling.

Essential Patterns

Always Use JSON Output

AI agents parse structured data, not ASCII tables. Always pass --format=json:

# Bad: returns an ASCII table that's hard to parse
wp llms course list

# Good: returns structured JSON
wp llms course list --format=json

Limit Fields to Reduce Context Size

LLM context windows are finite. Use --fields to request only what you need:

# All fields (verbose)
wp llms course get 42 --format=json

# Just what you need
wp llms course get 42 --fields=id,title,status --format=json

Use --porcelain for Create/Update Workflows

When creating or updating resources, --porcelain returns just the ID — useful for chaining:

# Create a course and capture its ID
COURSE_ID=$(wp llms course create --title="My Course" --status=draft --porcelain)

# Create a section in that course
SECTION_ID=$(wp llms section create --title="Getting Started" --parent_id=$COURSE_ID --porcelain)

# Create a lesson in that section
wp llms lesson create --title="Welcome" --parent_id=$SECTION_ID --status=publish --porcelain

Get Course Structure in One Call

Use wp llms course content to see the full outline:

wp llms course content 42 --format=json

Returns sections with their lessons, ordered by position. This is faster than listing sections and lessons separately.

Check Enrollment Status

# How many students in a course?
wp llms course enrollments 42 --format=count

# Who's enrolled?
wp llms course enrollments 42 --format=json

# Is student 5 enrolled in course 42?
wp llms students-enrollments list --student_id=5 --post_id=42 --format=json

Common Workflows

Create a Complete Course

# 1. Create the course
COURSE_ID=$(wp llms course create --title="Photography Basics" --status=draft --porcelain)

# 2. Create sections
S1=$(wp llms section create --title="Camera Fundamentals" --parent_id=$COURSE_ID --order=1 --porcelain)
S2=$(wp llms section create --title="Composition" --parent_id=$COURSE_ID --order=2 --porcelain)

# 3. Create lessons
wp llms lesson create --title="Aperture & Shutter Speed" --parent_id=$S1 --order=1 --status=publish --porcelain
wp llms lesson create --title="ISO & Exposure" --parent_id=$S1 --order=2 --status=publish --porcelain
wp llms lesson create --title="Rule of Thirds" --parent_id=$S2 --order=1 --status=publish --porcelain
wp llms lesson create --title="Leading Lines" --parent_id=$S2 --order=2 --status=publish --porcelain

# 4. Create an access plan
wp llms access-plan create --title="Full Access" --post_id=$COURSE_ID --price=49 --frequency=0

# 5. Publish
wp llms course update $COURSE_ID --status=publish

# 6. Verify
wp llms course content $COURSE_ID --format=json

Bulk Enrollment

# Enroll multiple students in a course
for STUDENT_ID in 5 12 28 45; do
  wp llms students-enrollments create --student_id=$STUDENT_ID --post_id=42
done

# Verify enrollments
wp llms course enrollments 42 --format=json

Progress Report

# Get all enrollments for a course
STUDENTS=$(wp llms course enrollments 42 --fields=student_id --format=json)

# Check each student's progress
wp llms students-progress list --post_id=42 --format=json

Find and Update

# Find a course by searching
wp llms course list --search="Python" --fields=id,title --format=json

# Update its title
wp llms course update 42 --title="Advanced Python Programming"

Resource Relationships

Understanding how LifterLMS resources relate to each other:

Course
├── Sections (ordered, belong to one course)
│   └── Lessons (ordered, belong to one section)
│       └── Quizzes (optional, attached to lessons)
├── Access Plans (pricing, one or more per course)
└── Enrollments (students enrolled in this course)
    └── Progress (per-student, per-course/lesson)

Membership
├── Access Plans (pricing)
├── Auto-enrollment courses (optional)
└── Enrollments (students)

Key IDs to track:

  • course_id / post_id: The course or membership ID
  • student_id: WordPress user ID
  • parent_id: Parent course (for sections) or parent section (for lessons)

Error Handling

CLI commands exit with standard exit codes:

  • 0 = success
  • 1 = error (with message on stderr)

Error messages are human-readable. For AI agents, check the exit code:

if wp llms course get 99999 --format=json 2>/dev/null; then
  echo "Course exists"
else
  echo "Course not found"
fi

CLI vs MCP Server

LifterLMS has both a CLI and an MCP server. When to use which:

Use CLI when... Use MCP when...
You have shell access to the WordPress server You're connecting remotely
Running Claude Code, Codex, or Cursor locally Using Claude Desktop or ChatGPT
Need to chain commands or write scripts Want conversational CRUD
Working in CI/CD pipelines Don't have SSH access

Both tools cover the same LifterLMS resources. The CLI talks to WordPress directly (internal REST requests). The MCP server talks over HTTP (external REST API with Application Passwords).