feature-engineering/task.toml at main · MMcollab-dotcom/feature-engineering · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
schema_version = "1.3"
artifacts = [{ source = "/app/submission", service = "mcp-server" }]

[task]
name = "mmcollab-dotcom/feature-engineering"
description = "Engineer causal features and submit one calibrated quantitative-ML trading strategy on a fixed public/hidden crypto split."
authors = [{ name = "EdotEnv", email = "michael.zhang@edotenv.com" }]
keywords = ["machine-learning-research", "quantitative-research", "quantitative-trading", "feature-engineering", "harbor"]

[metadata]
category = "machine-learning-research"
difficulty = "medium"
difficulty_explanation = "The agent performs quantitative ML and quantitative-trading research on fixed public rows, trains source-bound sklearn models, backtests candidates asynchronously, and submits one strategy."
suite = "feature-engineering"
family = "feature-engineering"

[verifier]
timeout_sec = 3600.0
environment_mode = "separate"

[agent]
timeout_sec = 14400.0

[environment]
build_timeout_sec = 1800.0
cpus = 4
memory_mb = 8192
storage_mb = 32768
gpus = 0
network_mode = "public"

[[environment.mcp_servers]]
name = "feature-engineering"
transport = "streamable-http"
url = "http://mcp-server:8000/mcp"