Run Harbor tasks from a standard Python image with runtime setup

#4
by burtenshaw HF Staff - opened
This view is limited to 50 files because it contains too many changes. See the raw diff here.
Files changed (50) hide show
  1. tasks/0000_324_324276_qa_3/environment/Dockerfile +0 -2
  2. tasks/0000_324_324276_qa_3/environment/requirements.txt +57 -0
  3. tasks/0000_324_324276_qa_3/task.toml +15 -7
  4. tasks/0000_369_369503_qa_1/environment/Dockerfile +0 -2
  5. tasks/0000_369_369503_qa_1/environment/requirements.txt +57 -0
  6. tasks/0000_369_369503_qa_1/task.toml +15 -7
  7. tasks/0000_455_455459_qa_4/environment/Dockerfile +0 -2
  8. tasks/0000_455_455459_qa_4/environment/requirements.txt +57 -0
  9. tasks/0000_455_455459_qa_4/task.toml +15 -7
  10. tasks/0000_465_465850_qa_5/environment/Dockerfile +0 -2
  11. tasks/0000_465_465850_qa_5/environment/requirements.txt +57 -0
  12. tasks/0000_465_465850_qa_5/task.toml +15 -7
  13. tasks/0000_526_526258_qa_2/environment/Dockerfile +0 -2
  14. tasks/0000_526_526258_qa_2/environment/requirements.txt +57 -0
  15. tasks/0000_526_526258_qa_2/task.toml +15 -7
  16. tasks/0000_539_539873_qa_3/environment/Dockerfile +0 -2
  17. tasks/0000_539_539873_qa_3/environment/requirements.txt +57 -0
  18. tasks/0000_539_539873_qa_3/task.toml +15 -7
  19. tasks/0000_582_582934_qa_4/environment/Dockerfile +0 -2
  20. tasks/0000_582_582934_qa_4/environment/requirements.txt +57 -0
  21. tasks/0000_582_582934_qa_4/task.toml +15 -7
  22. tasks/0000_587_587336_qa_5/environment/Dockerfile +0 -2
  23. tasks/0000_587_587336_qa_5/environment/requirements.txt +57 -0
  24. tasks/0000_587_587336_qa_5/task.toml +15 -7
  25. tasks/0000_641_641256_qa_1/environment/Dockerfile +0 -2
  26. tasks/0000_641_641256_qa_1/environment/requirements.txt +57 -0
  27. tasks/0000_641_641256_qa_1/task.toml +15 -7
  28. tasks/0000_656_656399_qa_2/environment/Dockerfile +0 -2
  29. tasks/0000_656_656399_qa_2/environment/requirements.txt +57 -0
  30. tasks/0000_656_656399_qa_2/task.toml +15 -7
  31. tasks/0000_767_767688_qa_4/environment/Dockerfile +0 -2
  32. tasks/0000_767_767688_qa_4/environment/requirements.txt +57 -0
  33. tasks/0000_767_767688_qa_4/task.toml +15 -7
  34. tasks/0000_780_780974_qa_4/environment/Dockerfile +0 -2
  35. tasks/0000_780_780974_qa_4/environment/requirements.txt +57 -0
  36. tasks/0000_780_780974_qa_4/task.toml +15 -7
  37. tasks/0000_804_804467_qa_1/environment/Dockerfile +0 -2
  38. tasks/0000_804_804467_qa_1/environment/requirements.txt +57 -0
  39. tasks/0000_804_804467_qa_1/task.toml +15 -7
  40. tasks/0000_804_804467_qa_3/environment/Dockerfile +0 -2
  41. tasks/0000_804_804467_qa_3/environment/requirements.txt +57 -0
  42. tasks/0000_804_804467_qa_3/task.toml +15 -7
  43. tasks/0000_806_806826_qa_3/environment/Dockerfile +0 -2
  44. tasks/0000_806_806826_qa_3/environment/requirements.txt +57 -0
  45. tasks/0000_806_806826_qa_3/task.toml +15 -7
  46. tasks/0000_849_849952_qa_4/environment/Dockerfile +0 -2
  47. tasks/0000_849_849952_qa_4/environment/requirements.txt +57 -0
  48. tasks/0000_849_849952_qa_4/task.toml +15 -7
  49. tasks/0000_886_886039_qa_2/environment/Dockerfile +0 -2
  50. tasks/0000_886_886039_qa_2/environment/requirements.txt +57 -0
tasks/0000_324_324276_qa_3/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_324_324276_qa_3/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_324_324276_qa_3/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "easy"
19
  difficulty_tier = "easy"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "freecodecamp__2016-new-coder-survey-"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "easy"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "freecodecamp__2016-new-coder-survey-"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_369_369503_qa_1/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_369_369503_qa_1/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_369_369503_qa_1/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 2
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "hugomathien__soccer"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 2
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "hugomathien__soccer"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_455_455459_qa_4/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_455_455459_qa_4/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_455_455459_qa_4/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "hard"
19
  difficulty_tier = "hard"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "abcsds__pokemon"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "hard"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "abcsds__pokemon"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_465_465850_qa_5/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_465_465850_qa_5/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_465_465850_qa_5/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "uciml__iris"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "uciml__iris"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_526_526258_qa_2/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_526_526258_qa_2/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_526_526258_qa_2/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "hard"
19
  difficulty_tier = "hard"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "marcomolina__water-consumption-in-a-median-size-city"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "hard"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "marcomolina__water-consumption-in-a-median-size-city"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_539_539873_qa_3/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_539_539873_qa_3/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_539_539873_qa_3/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "fbi-us__california-crime"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "fbi-us__california-crime"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_582_582934_qa_4/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_582_582934_qa_4/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_582_582934_qa_4/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "washingtonpost__police-shootings"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "washingtonpost__police-shootings"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_587_587336_qa_5/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_587_587336_qa_5/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_587_587336_qa_5/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "washingtonpost__police-shootings"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "washingtonpost__police-shootings"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_641_641256_qa_1/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_641_641256_qa_1/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_641_641256_qa_1/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "zed9941__top-500-indian-cities"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "zed9941__top-500-indian-cities"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_656_656399_qa_2/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_656_656399_qa_2/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_656_656399_qa_2/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "aljarah__xAPI-Edu-Data"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "aljarah__xAPI-Edu-Data"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_767_767688_qa_4/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_767_767688_qa_4/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_767_767688_qa_4/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "uciml__breast-cancer-wisconsin-data"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "uciml__breast-cancer-wisconsin-data"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_780_780974_qa_4/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_780_780974_qa_4/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_780_780974_qa_4/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "cbp__illegal-immigrants"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "cbp__illegal-immigrants"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_804_804467_qa_1/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_804_804467_qa_1/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_804_804467_qa_1/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "hard"
19
  difficulty_tier = "hard"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "uciml__mushroom-classification"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "hard"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "uciml__mushroom-classification"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_804_804467_qa_3/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_804_804467_qa_3/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_804_804467_qa_3/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "hard"
19
  difficulty_tier = "hard"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "uciml__mushroom-classification"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "hard"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "uciml__mushroom-classification"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_806_806826_qa_3/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_806_806826_qa_3/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_806_806826_qa_3/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "medium"
19
  difficulty_tier = "medium"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "groundhogclub__groundhog-day"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "medium"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "groundhogclub__groundhog-day"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_849_849952_qa_4/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_849_849952_qa_4/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0
tasks/0000_849_849952_qa_4/task.toml CHANGED
@@ -19,6 +19,8 @@ difficulty = "hard"
19
  difficulty_tier = "hard"
20
 
21
  [environment]
 
 
22
  build_timeout_sec = 600.0
23
  os = "linux"
24
  cpus = 1
@@ -28,18 +30,24 @@ gpus = 0
28
  allow_internet = true
29
  mcp_servers = []
30
 
31
- # Pre-agent hook: Harbor runs the command AFTER container start and BEFORE the
32
- # agent setup begins. We use it to pull this task's bucket prefix into
33
- # /home/user/input/. See environment/pull_bucket.py.
34
  [environment.healthcheck]
35
- command = "python3 /opt/pull_bucket.py && [ -n \"$(ls /home/user/input)\" ]"
 
 
 
 
 
 
36
  interval_sec = 2.0
37
- timeout_sec = 180.0
38
- start_period_sec = 5.0
39
  start_interval_sec = 2.0
40
- retries = 30
41
 
42
  [environment.env]
 
 
43
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
44
  BUCKET_PREFIX = "umichigan__world-religions"
45
  HF_TOKEN = "${HF_TOKEN}"
 
19
  difficulty_tier = "hard"
20
 
21
  [environment]
22
+ docker_image = "python@sha256:a0f57064cbdcd6285fa45f51237ab0b35ea8ebaed06a1f5973e89e19dfb86704"
23
+ workdir = "/workdir"
24
  build_timeout_sec = 600.0
25
  os = "linux"
26
  cpus = 1
 
30
  allow_internet = true
31
  mcp_servers = []
32
 
33
+ # Install dependencies and pull task data before agent setup.
 
 
34
  [environment.healthcheck]
35
+ command = '''
36
+ apt-get update &&
37
+ DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends ca-certificates curl git &&
38
+ python3 -m pip install --no-cache-dir -r /workdir/requirements.txt &&
39
+ python3 /workdir/pull_bucket.py &&
40
+ test -n "$(ls /home/user/input)"
41
+ '''
42
  interval_sec = 2.0
43
+ timeout_sec = 600.0
44
+ start_period_sec = 0.0
45
  start_interval_sec = 2.0
46
+ retries = 1
47
 
48
  [environment.env]
49
+ HF_XET_HIGH_PERFORMANCE = "1"
50
+ TAR_OPTIONS = "--no-same-owner"
51
  HF_BUCKET = "AdithyaSK/jupyter-agent-kaggle-all"
52
  BUCKET_PREFIX = "umichigan__world-religions"
53
  HF_TOKEN = "${HF_TOKEN}"
tasks/0000_886_886039_qa_2/environment/Dockerfile DELETED
@@ -1,2 +0,0 @@
1
- FROM docker.io/savatar101/env-data-agent-train:base
2
- WORKDIR /workdir
 
 
 
tasks/0000_886_886039_qa_2/environment/requirements.txt ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Captured with pip freeze from the existing upstream image:
2
+ # savatar101/env-data-agent-train@sha256:7b3f7e22780b431ce0d43717dcef6875cde84baeee824cc87e968374d62d3470
3
+ annotated-doc==0.0.4
4
+ annotated-types==0.7.0
5
+ antlr4-python3-runtime==4.13.2
6
+ anyio==4.14.0
7
+ certifi==2026.6.17
8
+ click==8.4.1
9
+ contourpy==1.3.3
10
+ cycler==0.12.1
11
+ distro==1.9.0
12
+ filelock==3.29.4
13
+ fonttools==4.63.0
14
+ fsspec==2026.6.0
15
+ h11==0.16.0
16
+ hf-xet==1.5.1
17
+ httpcore==1.0.9
18
+ httpx==0.28.1
19
+ huggingface_hub==1.19.0
20
+ idna==3.18
21
+ jiter==0.15.0
22
+ joblib==1.5.3
23
+ kiwisolver==1.5.0
24
+ latex2sympy2_extended==1.11.0
25
+ markdown-it-py==4.2.0
26
+ math-verify==0.9.0
27
+ matplotlib==3.11.0
28
+ mdurl==0.1.2
29
+ mpmath==1.3.0
30
+ narwhals==2.22.1
31
+ numpy==2.4.6
32
+ openai==2.43.0
33
+ packaging==26.2
34
+ pandas==3.0.3
35
+ patsy==1.0.2
36
+ pillow==12.2.0
37
+ pydantic==2.13.4
38
+ pydantic_core==2.46.4
39
+ Pygments==2.20.0
40
+ pyparsing==3.3.2
41
+ python-dateutil==2.9.0.post0
42
+ PyYAML==6.0.3
43
+ rich==15.0.0
44
+ scikit-learn==1.9.0
45
+ scipy==1.17.1
46
+ seaborn==0.13.2
47
+ shellingham==1.5.4
48
+ six==1.17.0
49
+ sniffio==1.3.1
50
+ statsmodels==0.14.6
51
+ sympy==1.14.0
52
+ tabulate==0.10.0
53
+ threadpoolctl==3.6.0
54
+ tqdm==4.68.3
55
+ typer==0.25.1
56
+ typing-inspection==0.4.2
57
+ typing_extensions==4.15.0