What Cato can do: demonstrated, not described.

tools: [
  "list_metrics",      // All 10 RC Charts metrics with descriptions
  "get_snapshot",      // MRR, active subs, trials, churn vs prior month
  "query_time_series"  // Any metric, any resolution, country filter
]

// Real API call with retry:
const url = `${RC_API_BASE}/charts/${metric}?${params}`;
const resp = await fetch(url, {
  headers: { Authorization: `Bearer ${RC_API_KEY}` }
});

def extract_content_insights(api_key, app_id) -> str:
    overview = get_subscription_overview(api_key, app_id)
    funnel = get_conversion_funnel(api_key, app_id)
    # Returns plain text for LLM grounding:
    # "Active subscriptions: 3,680"
    # "MRR: $51,600"
    # "Trial conversion rate: 10.9%"
    # "Monthly churn: 4.3%"

@bot.event
async def on_message(message):
    if message.channel.id not in MONITORED_CHANNELS:
        return
    # Keyword classification: bug, feature, question
    signal_type = classify_signal(message.content)
    db.insert("community_signals", {
        "platform": "discord",
        "signal_type": signal_type,
        "body": message.content[:2000],
        "author": str(message.author),
        "source_url": message.jump_url,
    })

def post_thread(text: str, handle: str, password: str):
    client = atproto.Client()
    client.login(handle, password)
    chunks = _split_thread(text, max_chars=300)
    parent = None
    for i, chunk in enumerate(chunks):
        record = _create_record(client, chunk, parent)
        if i == 0:
            parent = {"uri": record.uri, "cid": record.cid}