AI Workflows

Agenda Panda works with any model, but the most reliable workflow is still: gather real brand context, generate or edit a reviewable calendar.json, dry-run the changes, then open the dashboard before applying them.

Default calendar workflow

# Gather context first
ap context --json > context.json

# Ask your model or agent to create or update calendar.json
claude "Use context.json to write next week's launch posts as CalendarItem JSON in calendar.json"

# Preview before applying
ap calendar sync --file calendar.json --dry-run
ap dashboard

Editing an existing calendar

When the schedule already exists, pull first instead of reconstructing it:

ap calendar pull --file calendar.json
ap context --json > context.json
claude "Use context.json and calendar.json to update next week's LinkedIn and X posts"
ap calendar sync --file calendar.json --dry-run
ap dashboard

One-off posts

For a single post, ap post is still the right tool:

ap context --json | claude "Write one LinkedIn post in our voice"
ap post create --content "Paste the approved caption here" --schedule 2026-03-15T14:00:00Z --connection linkedin

Net-new visuals

Generate or edit media first, then reuse the hosted result:

ap media generate "editorial product hero shot, soft daylight, no people" --aspect 4:5
ap post "Launch visual" --media-id abc12345
ap media edit abc12345 "same shot, darker background, warmer light"

Brand Context

Use ap context --json to give your AI model real context about your brand:

ap context --json | claude "Write a launch announcement in our voice"

This pipes your soul file, art direction, channel info, and media workflow rules to the model.

Agent prompt pattern

Good agent prompts ask for a whole reviewable workflow, not a loop of one-off posts. For example:

Use context.json to build calendar.json for next week's launch posts, preview it with ap calendar sync --file calendar.json --dry-run, and open ap dashboard for review. Do not create the calendar one post at a time.

SKILL.md

Agenda Panda ships a SKILL.md reference file that AI coding agents (Claude Code, Cursor, Windsurf) can read to understand how to use the CLI.

cat SKILL.md

Your AI agent instantly knows the commands, JSON format, and the preferred workflow: gather context, generate media when needed, build calendar.json, dry-run sync, review in the dashboard, then apply.