rlm-execute - SKILL.md Agent Skill

name: rlm-execute description: Execute Python code in Daytona sandboxes with durable volume persistence. Use when running code in a Daytona sandbox, processing data with stateful execution, or persisting results across sessions.

RLM Execute

Execute Python code in Daytona sandboxes with volume persistence.

Basic Execution

from fleet_rlm.runtime.config import configure_planner_from_env
from fleet_rlm.integrations.providers.daytona.interpreter import DaytonaInterpreter

configure_planner_from_env()

interp = DaytonaInterpreter(
    repo_url="https://github.com/your-org/your-repo",
    timeout=600,
)
interp.start()
try:
    result = interp.execute('''
import math
SUBMIT(answer=math.factorial(15))
''')
    print(result.answer)  # 1307674368000
finally:
    interp.shutdown()

With Durable Volume Persistence

Data written to /home/daytona/memory/ persists across sessions on the mounted Daytona volume.

interp = DaytonaInterpreter(
    repo_url="https://github.com/your-org/your-repo",
    volume_name="my-data",
    timeout=600,
)
interp.start()
try:
    # Write data
    interp.execute('''
import json, os
os.makedirs('/home/daytona/memory/artifacts', exist_ok=True)
data = {"processed": True, "count": 42}
with open('/home/daytona/memory/artifacts/results.json', 'w') as f:
    json.dump(data, f)
SUBMIT(status="saved")
''')

    # Read it back in the same session (or a later one)
    result = interp.execute('''
import json
with open('/home/daytona/memory/artifacts/results.json') as f:
    data = json.load(f)
SUBMIT(data=data)
''')
    print(result.data)  # {"processed": True, "count": 42}
finally:
    interp.shutdown()

Execute a Local File

# Read local file, execute in sandbox
code = open("scripts/analysis.py").read()
interp.start()
try:
    result = interp.execute(code)
finally:
    interp.shutdown()

Execution Patterns

Data Processing Pipeline

interp = DaytonaInterpreter(
    repo_url="https://github.com/your-org/your-repo",
    volume_name="pipeline",
    timeout=600,
)
interp.start()
try:
    # Step 1: Generate data
    interp.execute('''
import json, os
os.makedirs('/home/daytona/memory/buffers', exist_ok=True)
data = [{"id": i, "value": i**2} for i in range(100)]
with open('/home/daytona/memory/buffers/raw.json', 'w') as f:
    json.dump(data, f)
SUBMIT(count=len(data))
''')

    # Step 2: Process (same sandbox)
    result = interp.execute('''
import json
with open('/home/daytona/memory/buffers/raw.json') as f:
    data = json.load(f)
filtered = [d for d in data if d["value"] > 50]
SUBMIT(filtered_count=len(filtered))
''')
    print(result.filtered_count)
finally:
    interp.shutdown()

Multi-Step with Buffers

add_buffer(name, value) / get_buffer(name) are injected by the sandbox driver (runtime/execution/sandbox_assets.py) and persist named lists across execute() calls in the same sandbox session:

interp.start()
try:
    # Buffers persist across execute() calls within same sandbox
    interp.execute('add_buffer("findings", "Step 1: setup complete")')
    interp.execute('add_buffer("findings", "Step 2: data loaded")')
    result = interp.execute('''
items = get_buffer("findings")
SUBMIT(log=items)
''')
    print(result.log)  # ["Step 1: setup complete", "Step 2: data loaded"]
finally:
    interp.shutdown()

Key Points

Access results via result.field_name (dot notation), not result["field"]
/home/daytona/memory/buffers/ is the canonical path for intermediate pipeline data
Data in /home/daytona/memory/ persists across sessions when using volume_name
Buffers (add_buffer/get_buffer) persist within a single sandbox session only
Always call interp.shutdown() in a finally block
Set timeout appropriately for long-running tasks; default is 900s

Multi-Step Pipeline (Canonical Pattern)

Use /home/daytona/memory/buffers/ for intermediate data in multi-step pipelines:

interp = DaytonaInterpreter(
    repo_url="https://github.com/your-org/your-repo",
    volume_name="pipeline-run",
    timeout=600,
)
interp.start()
try:
    # Step 1: generate and persist intermediate data
    interp.execute('''
import json, os
os.makedirs('/home/daytona/memory/buffers', exist_ok=True)
raw = [{"id": i, "score": i * 1.5} for i in range(50)]
with open('/home/daytona/memory/buffers/step1_raw.json', 'w') as f:
    json.dump(raw, f)
SUBMIT(count=len(raw))
''')

    # Step 2: load from buffers, process, write final result
    result = interp.execute('''
import json, os
with open('/home/daytona/memory/buffers/step1_raw.json') as f:
    raw = json.load(f)
filtered = [r for r in raw if r["score"] > 30]
os.makedirs('/home/daytona/memory/artifacts', exist_ok=True)
with open('/home/daytona/memory/artifacts/final.json', 'w') as f:
    json.dump(filtered, f)
SUBMIT(filtered_count=len(filtered))
''')
    print(result.filtered_count)
finally:
    interp.shutdown()

Troubleshooting

See rlm-debug for runtime failures and daytona-runtime for Daytona-specific execution rules.