LF Security Corpus: Dataset Poisoning & Eval Integrity
Collection
Synthetic poisoning, duplication, trigger, contamination and negative-control datasets. • 12 items • Updated
Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
artifacts: list<item: string>
child 0, item: string
family: string
note: string
repo_name: string
attack_surface: list<item: string>
child 0, item: string
corpus_id: string
safe_fixture: bool
control_type: string
ground_truth: string
related_cases: list<item: null>
child 0, item: null
difficulty: string
expected_decision: string
schema: int64
transformations: list<item: null>
child 0, item: null
layerfault_rule_expectations: struct<candidate_rules: list<item: null>, expected_rules: list<item: null>, must_not_rules: list<ite (... 9 chars omitted)
child 0, candidate_rules: list<item: null>
child 0, item: null
child 1, expected_rules: list<item: null>
child 0, item: null
child 2, must_not_rules: list<item: null>
child 0, item: null
techniques: list<item: string>
child 0, item: string
oracle_ids: list<item: string>
child 0, item: string
severity: string
to
{'attack_surface': List(Value('string')), 'control_type': Value('string'), 'corpus_id': Value('string'), 'difficulty': Value('string'), 'expected_decision': Value('string'), 'ground_truth': Value('string'), 'layerfault_rule_expectations': {'candidate_rules': List(Value('null')), 'expected_rules': List(Value('null')), 'must_not_rules': List(Value('null'))}, 'oracle_ids': List(Value('string')), 'related_cases': List(Value('null')), 'repo_name': Value('string'), 'safe_fixture': Value('bool'), 'schema': Value('int64'), 'severity': Value('string'), 'techniques': List(Value('string')), 'transformations': List(Value('null'))}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
artifacts: list<item: string>
child 0, item: string
family: string
note: string
repo_name: string
attack_surface: list<item: string>
child 0, item: string
corpus_id: string
safe_fixture: bool
control_type: string
ground_truth: string
related_cases: list<item: null>
child 0, item: null
difficulty: string
expected_decision: string
schema: int64
transformations: list<item: null>
child 0, item: null
layerfault_rule_expectations: struct<candidate_rules: list<item: null>, expected_rules: list<item: null>, must_not_rules: list<ite (... 9 chars omitted)
child 0, candidate_rules: list<item: null>
child 0, item: null
child 1, expected_rules: list<item: null>
child 0, item: null
child 2, must_not_rules: list<item: null>
child 0, item: null
techniques: list<item: string>
child 0, item: string
oracle_ids: list<item: string>
child 0, item: string
severity: string
to
{'attack_surface': List(Value('string')), 'control_type': Value('string'), 'corpus_id': Value('string'), 'difficulty': Value('string'), 'expected_decision': Value('string'), 'ground_truth': Value('string'), 'layerfault_rule_expectations': {'candidate_rules': List(Value('null')), 'expected_rules': List(Value('null')), 'must_not_rules': List(Value('null'))}, 'oracle_ids': List(Value('string')), 'related_cases': List(Value('null')), 'repo_name': Value('string'), 'safe_fixture': Value('bool'), 'schema': Value('int64'), 'severity': Value('string'), 'techniques': List(Value('string')), 'transformations': List(Value('null'))}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
dataset-label-flip-minority
SECURITY TEST ARTIFACT: DO NOT USE AS A PRODUCTION MODEL
This repository is part of the Layerfault synthetic security corpus. It is deliberately constructed to contain security-relevant characteristics for scanner testing.
Corpus ID: LF-CH-DATA-0009
Purpose
Dataset label flip minority.
Direct expected Layerfault rules
- None; this repository is a control/comparison input.
Candidate rules
These are deliberately plausible targets that remain marked as candidates until the exact Layerfault build used for certification confirms them.
- None
Negative-control rules
These should remain silent for this corpus item.
- None
Safety
The corpus uses fake secrets, loopback/.invalid network destinations, harmless marker output,
and synthetic model behavior only. It is intended for static scanning and isolated security testing.
Challenge classification
- Severity: high
- Difficulty: compound
- Expected admission decision: BLOCK
- Control type: positive
- Attack surface: dataset-poisoning
- Techniques: dataset
- Transformations: none
Ground-truth oracle IDs
LF-ORACLE-DATA-0009
These oracle IDs describe synthetic ground truth. They do not claim that a matching Layerfault detector already exists. A challenge may intentionally expose a scanner blind spot and remain unmapped until the detector is implemented.
- Downloads last month
- 28