Autopilot quickstart

from pathlib import Path
import sys

# Resolve the repo root whether this cell is running from ./notebooks
# or from the repo root — mirrors the existing 01–36 notebook series.
ROOT = (
    Path.cwd().resolve().parent
    if Path.cwd().name == "notebooks"
    else Path.cwd().resolve()
)
if str(ROOT) not in sys.path:
    sys.path.insert(0, str(ROOT))

from examples.run_multi_tool_agent import build_llm_from_env

# Bedrock Llama 4 Scout is the default model this library ships with
# (see `shipit_agent/config.py`). No model name required — the helper
# reads `AWS_REGION` / credentials from env and returns an adapter that
# works with every Autopilot / Agent class.
llm = build_llm_from_env("bedrock")
print("Bedrock LLM ready:", type(llm).__name__)

from shipit_agent.deep import Goal

goal = Goal(
    objective="Summarize the three most common Python dict gotchas and show a working snippet for each.",
    success_criteria=[
        "At least 3 gotchas explained in prose",
        "Each gotcha has a Python snippet that reproduces it",
        "A one-line fix or avoidance is shown for each",
    ],
)
goal

from shipit_agent.autopilot import Autopilot, BudgetPolicy, default_heartbeat_stderr

autopilot = Autopilot(
    llm=llm,
    goal=goal,
    budget=BudgetPolicy(
        max_seconds=600,  # 10 min cap for the demo
        max_tool_calls=20,
        max_iterations=8,
        max_tokens=300_000,
        max_dollars=2.0,
    ),
    heartbeat_every_seconds=30,
    on_heartbeat=default_heartbeat_stderr,
)
print("Autopilot ready — budget:", autopilot.budget)

result = autopilot.run(run_id="py-dict-gotchas-v1")

print(f"status:       {result.status}")
print(f"iterations:   {result.iterations}")
print(
    f"criteria met: {sum(1 for c in result.criteria_met if c)} / {len(result.criteria_met)}"
)
print(f"halt reason:  {result.halt_reason}")
print()
print(result.output[:800])

import json
from pathlib import Path

cp = Path.home() / ".shipit_agent" / "checkpoints" / "py-dict-gotchas-v1.json"
if cp.exists():
    data = json.loads(cp.read_text())
    print(json.dumps({k: data[k] for k in ("run_id", "iterations", "usage")}, indent=2))
else:
    print("(no checkpoint — run cell 3 first)")