# Demo: kj audit --agent-readiness — score third-party repos for AI-agent compatibility
# Length: ~1 minute. No LLM calls. Reproducible against any tagged release.
#
# Setup: cd into the karajan-code repo (or any other) BEFORE starting asciinema.
# The script below lists the EXACT commands to type, with optional pauses
# the user can read while typing.

# 1. Show the score for a repo that's NOT prepared (third-party).
#    Recommended: clone https://github.com/expressjs/express somewhere
#    (it has no llms.txt → low score, big contrast vs Karajan's 100/100).
#    Replace the path below with wherever you cloned it.
clear
ls ~/oss/express
kj audit --agent-readiness --path ~/oss/express

# (Pause ~3s after the score appears — the audience needs to read
# the per-check breakdown and the top-fixes list.)

# 2. Now the punchline: same command on Karajan Code itself.
clear
kj audit --agent-readiness

# (The score line says 100/100. Pause ~3s.)

# 3. Show that it's machine-readable too — pipe to jq for CI use.
kj audit --agent-readiness --json | jq '.score'

# 4. Optional: hint at what's behind the score by tailing llms.txt.
head -20 llms.txt
ls docs/agents/SKILL.kj-*.md

# Done. Hit Ctrl+D or type `exit` to stop the recording.
