MCPcopy Create free account

hub / github.com/NVIDIA/SkillEvaluator / functions

Functions24,741 in github.com/NVIDIA/SkillEvaluator

↓ 1 callersFunction_load_existing_cases
(output_path: Path)
src/skillevaluator/tier3/generate_dataset.py:1671
↓ 1 callersFunction_load_existing_key
()
src/skillevaluator/tier3/output_provenance.py:389
↓ 1 callersFunction_load_frontmatter
Bounded YAML frontmatter, with Claude Code's lenient retry when strict YAML rejects it. A key written twice keeps its last value, as in Claude Co
src/skillevaluator/plugin_markdown.py:72
↓ 1 callersFunction_load_json
(path: Path)
src/skillevaluator/tier3/dataset_utils.py:190
↓ 1 callersFunction_load_json
(env_dir: Path, rel: str, cache: dict[str, Any])
src/skillevaluator/tier3/loose_parts.py:1382
↓ 1 callersFunction_load_jsonl
(path: Path)
src/skillevaluator/tier3/dataset_utils.py:175
↓ 1 callersFunction_load_or_create_key
()
src/skillevaluator/tier3/output_provenance.py:449
↓ 1 callersFunction_load_path_cached
(path: str, fingerprint: tuple[int, int, int] | None)
src/skillevaluator/reporting/dashboard_app.py:108
↓ 1 callersFunction_load_payload
(reports: Path, relative: str)
ci/benchmark_reports.py:52
↓ 1 callersFunction_load_report
(path: Path)
ci/run_tier1.py:213
↓ 1 callersFunction_load_report_file
( path: Path, ancestors: frozenset[Path] = frozenset(), budget: _CatalogBudget | None = None )
src/skillevaluator/reporting/dashboard_data.py:755
↓ 1 callersMethod_load_script
(self, rel: PurePosixPath)
src/skillevaluator/plugin_component_risk.py:3297
↓ 1 callersFunction_load_state
Load state from markdown file, returning (state, error).
src/skillevaluator/tier3/reference_skills/task-list/scripts/tasks.py:45
↓ 1 callersFunction_load_template
()
scripts/script_invocation_differential.py:69
↓ 1 callersFunction_load_template
()
tests/tier3/test_negative_control_invocation_evidence.py:19
↓ 1 callersFunction_load_template
()
tests/tier3/test_script_invocation_evidence.py:22
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_codex_tool_call_normalization.py:36
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_security_finding_evidence.py:38
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_github_token_redaction.py:21
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_network_client_mentions.py:33
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_apply_patch_security.py:25
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_judge_parse_robustness.py:47
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_harbor_template_secret_patterns.py:28
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_issue192_security_false_positives.py:42
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_security_home_paths.py:26
↓ 1 callersFunction_load_template_module
()
tests/tier3/test_judge_result_semantics.py:43
↓ 1 callersFunction_load_trial_rewards
Load bounded, contained with-skill rewards for one agent.
src/skillevaluator/tier3/harbor/report.py:113
↓ 1 callersFunction_load_views
(projection: _RuntimeProjection, native: dict, decisions: dict)
src/skillevaluator/reporting/plugin_runtime.py:147
↓ 1 callersFunction_load_yaml
(path: Path)
src/skillevaluator/tier3/dataset_utils.py:185
↓ 1 callersMethod_load_yaml
Parse the manifest YAML; record a finding and return None on failure.
src/skillevaluator/validators/plugin_schema.py:136
↓ 1 callersFunction_loaded
Return :class:`PIIWaiver` fields for a waiver built without the loader.
tests/validators/test_pii_waivers.py:46
↓ 1 callersFunction_loaded_dependency_skill
The dependency folder of a ``SKILL.md`` a client loads from a skills folder, or ``None``. ``parts`` is the manifest path relative to the skills f
src/skillevaluator/plugin_member_paths.py:229
↓ 1 callersFunction_local_reference_escapes_repo
(uses: str)
tests/test_ci_workflows.py:491
↓ 1 callersFunction_local_skill_dir
The evaluated skill's directory on this machine, when it is known. An explicit ``--skill`` is already resolved. An inferred skill is only the
src/skillevaluator/evaluation/tier3_feedback_chat.py:567
↓ 1 callersMethod_locate_manifest
Return ``(manifest_path, manifest_type)`` with bundle precedence.
src/skillevaluator/validators/plugin_schema.py:106
↓ 1 callersFunction_located_issue
Return an issue's normalized location and all its normalized report paths. The report paths are None when the issue's occurrences are malformed.
src/skillevaluator/validators/skillspector_scope.py:588
↓ 1 callersMethod_location_end
(self, i: int)
src/skillevaluator/evaluation/dataset_case_removal.py:305
↓ 1 callersMethod_location_paths
Return ``file_path`` and the shorter paths a message may name it by. A folder run prefixes each skill's ``file_path`` with ``[skill] ``
src/skillevaluator/models/result.py:141
↓ 1 callersFunction_logical_lines
``run`` split into lines, with backslash continuations joined back up.
tests/test_ci_workflows.py:645
↓ 1 callersFunction_logical_reward_mean
Average a persisted reward field once per logical Harbor trial.
src/skillevaluator/evaluation/tier3_report.py:5157
↓ 1 callersFunction_looks_like_path
A file path (``/run/secrets/db``, ``./key.pem``, ``~/.config/x``), not a token that starts with ``/``. A base64 secret starts with ``/`` about on
src/skillevaluator/validators/url_policy.py:378
↓ 1 callersMethod_looks_like_value
(self, raw: str)
src/skillevaluator/cli.py:449
↓ 1 callersFunction_lookup
(args: Any, name: str)
src/skillevaluator/tier3/eval_core/plugin_signals.py:3800
↓ 1 callersFunction_loop_header_is_empty
``for f in; do`` with nothing after ``in``: the loop runs zero times. Its variable keeps whatever it held, and its body never runs, so the bo
src/skillevaluator/tier3/eval_core/script_invocation.py:1993
↓ 1 callersFunction_loop_header_is_unresolved
Bind a loop variable where the header settles it, else say whether it matters. ``for f in run.py; do python $f; done`` gives ``f`` exactly one va
src/skillevaluator/tier3/eval_core/script_invocation.py:2010
↓ 1 callersFunction_loopback
(statuses: list[int])
tests/test_transport_rules.py:141
↓ 1 callersFunction_lsp_command_findings
The MCP stdio command-form checks for one LSP server's ``command`` and ``args``, as ``plugin_lsp_*``. Claude Code starts an LSP server like a std
src/skillevaluator/plugin_config_risk.py:69
↓ 1 callersFunction_lsp_env_findings
The MCP env checks for an LSP server's ``env``: TLS verification off and inline credentials (CRITICAL).
src/skillevaluator/plugin_config_risk.py:127
↓ 1 callersFunction_lsp_extension_index
File extension -> the one LSP server declared in this arm that serves it. An extension two declared servers claim is left out: the call cannot sa
src/skillevaluator/tier3/eval_core/plugin_signals.py:1514
↓ 1 callersFunction_lsp_ident
The plugin LSP server a Claude Code ``LSP`` call reached, from its file's extension, or ``None``.
src/skillevaluator/tier3/eval_core/plugin_signals.py:1528
↓ 1 callersFunction_lsp_wording
MCP check text reworded for an LSP server.
src/skillevaluator/plugin_config_risk.py:118
↓ 1 callersFunction_manifest_file
A regular single-link manifest candidate, ``None`` when it vanished; anything else fails closed.
src/skillevaluator/plugin_manifest.py:440
↓ 1 callersFunction_manifest_operands
Operands this command really reads as file text: reader verbs only. ``cp``/``mv`` sources, ``<`` input to a non-reader (``grep x < SKILL.md``),
src/skillevaluator/tier3/eval_core/plugin_signals.py:2617
↓ 1 callersFunction_manifest_reference_text
Keep manifest runtime references while excluding descriptive metadata.
src/skillevaluator/validators/skillspector_scope.py:780
↓ 1 callersFunction_manifests
(paths: list[str])
ci/select_changed_skills.py:193
↓ 1 callersMethod_map_bandit_severity
Map Bandit severity string to Severity enum.
src/skillevaluator/validators/code_risk.py:303
↓ 1 callersFunction_mapped_arguments
Read a mapped call's argument: the element itself, or an object literal naming it once. Returns ``(shorthand_key, literal_properties)``. ``shorth
src/skillevaluator/tier3/eval_core/codex_tool_call_normalizer.py:1093
↓ 1 callersFunction_mapped_tool_calls
Read ``<array>.map(<callback>)`` and return one call per literal element, in array order.
src/skillevaluator/tier3/eval_core/codex_tool_call_normalizer.py:1128
↓ 1 callersFunction_mapping
(value: object)
src/skillevaluator/reporting/markdown.py:137
↓ 1 callersFunction_mark_failed
(target: dict[str, Any], errors: list[str])
src/skillevaluator/evaluation/tier3_report.py:629
↓ 1 callersFunction_mark_publication_target_conflict
Remove contradictory target claims before persisting a conflict.
src/skillevaluator/publication_identity.py:768
↓ 1 callersFunction_mark_reward_collection_failure
(reward: dict[str, Any], reason: str)
src/skillevaluator/tier3/harbor/collector.py:4462
↓ 1 callersFunction_mark_work_limit_incomplete
Record a tripped work limit as incomplete coverage, not a duplicate finding. The scan stopped before duplicate analysis finished, so it can neith
src/skillevaluator/deduplication/intra_skill/intra_skill_validator.py:142
↓ 1 callersFunction_markdown_blocks
Keep each generated table contiguous and atomic under display limits.
ci/report_views.py:234
↓ 1 callersFunction_markdown_code_text
Flatten untrusted metadata for a code span without obscuring paths.
src/skillevaluator/reporting/markdown.py:127
↓ 1 callersFunction_markdown_comment_line_numbers
Skip YAML frontmatter comments and fenced-code comments, not ATX headings.
src/skillevaluator/validators/security.py:431
↓ 1 callersFunction_markdown_cost
The cost row of an agent, command, or output style (a forced style is resolved per view).
src/skillevaluator/plugin_context_cost.py:363
↓ 1 callersFunction_markdown_fence_comment_line_numbers
(lines: list[str], start: int)
src/skillevaluator/validators/security.py:412
↓ 1 callersFunction_markdown_frontmatter_end
(lines: list[str])
src/skillevaluator/validators/security.py:403
↓ 1 callersFunction_markdown_safe_url
Return an HTTP(S) URL that cannot terminate a Markdown destination.
src/skillevaluator/reporting/markdown.py:150
↓ 1 callersFunction_markdown_table_cell
Return one safe physical Markdown table cell.
src/skillevaluator/reporting/markdown.py:73
↓ 1 callersFunction_marker
(name: str)
scripts/script_invocation_differential.py:121
↓ 1 callersFunction_marketplace_source_dir
The folder a relative marketplace ``source`` names (Claude Code resolves it from the marketplace root).
src/skillevaluator/plugin_manifest.py:621
↓ 1 callersMethod_mask
(self)
src/skillevaluator/tier3/eval_core/checks.py:4583
↓ 1 callersFunction_mask_argument_hint
Mask the short CLI placeholders in a string ``argument-hint`` for the XML tag check. Control tags are checked on the raw value first: masking an
src/skillevaluator/validators/quality_score.py:152
↓ 1 callersFunction_mask_argv_credentials
Mask the list element after a quoted credential flag (["tool", "--password", "VALUE"]).
src/skillevaluator/tier3/eval_core/checks.py:4647
↓ 1 callersFunction_mask_authorization_headers
Mask Authorization header values: to the closing quote when quoted, else one word (two after a scheme).
src/skillevaluator/tier3/eval_core/checks.py:4264
↓ 1 callersFunction_mask_config_set_values
Mask the value of each config set command in *text* whose key names a credential.
src/skillevaluator/tier3/eval_core/checks.py:4743
↓ 1 callersFunction_mask_credential_assignments
Mask the value of each assignment whose name holds a credential word ("_authToken=...", "api_key: ...").
src/skillevaluator/tier3/eval_core/checks.py:4989
↓ 1 callersFunction_mask_credential_flags
Mask the value of each flag whose name holds a credential word ("--client-secret VALUE").
src/skillevaluator/tier3/eval_core/checks.py:4293
↓ 1 callersFunction_mask_digest_headers
Mask Digest Authorization values: to the first unescaped quote inside double quotes, else as judge evidence.
src/skillevaluator/tier3/eval_core/checks.py:4241
↓ 1 callersFunction_mask_finding_evidence_credentials
Mask supported credential forms, also in text escaped again, without imposing a caller's evidence budget.
src/skillevaluator/tier3/eval_core/checks.py:3989
↓ 1 callersFunction_mask_finding_evidence_credentials_as_written
Mask supported credential forms without imposing a caller's evidence budget.
src/skillevaluator/tier3/eval_core/checks.py:4217
↓ 1 callersFunction_mask_judge_evidence_line
Mask credentials in the text between two escapes, which is not trimmed.
src/skillevaluator/tier3/harbor/templates/eval.py:2899
↓ 1 callersFunction_mask_judge_evidence_text
Mask credentials in *text* read as one piece.
src/skillevaluator/tier3/harbor/templates/eval.py:2888
↓ 1 callersFunction_mask_judge_evidence_value_sites
Mask the values of each text in *sites* that *text* holds between a marker and the next escape or marker.
src/skillevaluator/tier3/harbor/templates/eval.py:2990
↓ 1 callersFunction_mask_non_assignment_equals
Replace every ``=`` that does not make its word a shell assignment. Tokens lose their quoting, but the shell only reads ``NAME=`` as an assig
src/skillevaluator/tier3/eval_core/checks.py:7928
↓ 1 callersFunction_mask_object
(value, redactor: Redactor)
ci/diagnostics.py:80
↓ 1 callersFunction_mask_skill_name
Replace whole literal identities without compiling contributor-controlled regexes.
src/skillevaluator/tier3/evals_spec.py:823
↓ 1 callersFunction_mask_spans
Return *masked*, a masker's output for *text*, with the parts of *spans* (of *text*) that it keeps masked.
src/skillevaluator/tier3/eval_core/checks.py:4167
↓ 1 callersFunction_mask_unescaped_views
Return *masked*, which *mask* wrote for *text*, also masking what *mask* masks in *text* unescaped. *text* is read with one level of escapes remo
src/skillevaluator/tier3/eval_core/checks.py:3998
↓ 1 callersFunction_masked_spans
Return the spans of *text* that *masked*, a masker's output for it, masks in any placement of what it keeps. Where the text kept on both sides of
src/skillevaluator/tier3/eval_core/checks.py:4094
↓ 1 callersFunction_masking_hides_text
Return whether masking left a ``>`` that could hide text inside a group. A ``>`` that closes no group (``<cmd>file>``) reads as a redirect. Insid
src/skillevaluator/validators/quality_score.py:113
↓ 1 callersFunction_match_is_inside_ranges
(match: re.Match[str], ranges: Iterable[tuple[int, int]])
scripts/check_oss_boundary.py:524
↓ 1 callersFunction_match_skill_name
(observed: str, expected: str, *, fuzzy: bool = False)
src/skillevaluator/tier3/eval_core/checks.py:3244
↓ 1 callersFunction_matcher_scope
:func:`matcher_scope` of a stripped matcher of at most ``MAX_MATCHER_CHARS`` characters.
src/skillevaluator/plugin_component_risk.py:1579
↓ 1 callersFunction_matcher_selects
Whether a tool matcher selects ``tool``, read the same way as :func:`matcher_scope`. It fails closed: a matcher the evaluator cannot decide selec
src/skillevaluator/plugin_component_risk.py:1661
↓ 1 callersFunction_matches
(rows: list[dict[str, str]])
src/skillevaluator/tier3/harbor/collector.py:7534
↓ 1 callersFunction_matching_baseline_tolerance
Return an agent's baseline-timeout record when it still matches the on-disk arms. The collector writes the record only for an agent it kept score
src/skillevaluator/tier3/harbor/report_data.py:669
↓ 1 callersFunction_matching_delimiters
(tokens: list[tuple[str, str]], opening: str, closing: str)
src/skillevaluator/tier3/eval_core/codex_tool_call_normalizer.py:696
← previousnext →6,801–6,900 of 24,741, ranked by callers