Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/NVIDIA/SkillEvaluator
/ functions
Functions
24,741 in github.com/NVIDIA/SkillEvaluator
⨍
Functions
24,741
◇
Types & classes
1,292
↳
Endpoints
703
↓ 3 callers
Function
c
(t,e)
src/skillevaluator/reporting/assets/chart.umd.min.js:7
↓ 3 callers
Function
canonical_agent_name
Return the Harbor agent identifier for a public CLI/config name.
src/skillevaluator/tier3/harbor/__init__.py:78
↓ 3 callers
Function
ce
(t,e,i)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Function
changed_paths
Return both sides of every path changed from the DIFF_BASE_SHA merge base to head.
ci/select_changed_skills.py:308
↓ 3 callers
Function
check_error_recovery
Detect error-retry patterns and attribute skill, agent, or environment faults. Scans bash/execute tool calls for commands that failed (non-zero e
src/skillevaluator/tier3/eval_core/checks.py:7399
↓ 3 callers
Function
check_quality
(report: dict, expected: dict)
scripts/demo_starter.py:53
↓ 3 callers
Function
child_process_env
Return a child environment without inherited observability exporters.
src/skillevaluator/utils/process_environment.py:12
↓ 3 callers
Function
co
(t,e,i,s)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Function
codex_steps
A user prompt, a model call that runs one tool 2 s later, and a final answer.
tests/reporting/test_resource_comparison.py:73
↓ 3 callers
Function
coerce_gke_opt_in_flag
Return True when an opt-in flag value represents an explicit truthy setting.
src/skillevaluator/tier3/harbor/gke_environment.py:361
↓ 3 callers
Function
compact_declarative_checks
Cap each arm's per-check rows, keeping exact counts and recording what was left out. Failed and not-checked rows are kept ahead of passed ones. T
src/skillevaluator/tier3/declarative_coverage.py:77
↓ 3 callers
Function
completeness_record
Project only diagnostic fields; do not copy arbitrary scanner payloads.
src/skillevaluator/skillspector_diagnostics.py:30
↓ 3 callers
Method
components
(self)
src/skillevaluator/deduplication/intra_skill/semantic_clustering.py:60
↓ 3 callers
Function
compute_adc_job_timeout
Compute the maximum safe Harbor job duration for an ADC access token.
src/skillevaluator/provider_config.py:290
↓ 3 callers
Function
condition_eligibility
Validate a retained arm, or the runner's not-yet-published collection.
src/skillevaluator/tier3/baseline_store.py:245
↓ 3 callers
Method
consume
(self, chunk: bytes)
src/skillevaluator/tier3/harbor/stream_redaction.py:32
↓ 3 callers
Function
continue_on_failure_scope
Toggle continue-on-failure behavior for the duration of the scope. While *enabled* is True, :meth:`ValidatorBase._validate_folder_or_skill` k
src/skillevaluator/validators/base.py:58
↓ 3 callers
Function
count
(name: str)
src/skillevaluator/evaluation/tier3_report.py:3458
↓ 3 callers
Function
coverage_notices
Return one notice per incomplete arm, in agent order, for the canonical payload.
src/skillevaluator/reporting/tier3_coverage.py:124
↓ 3 callers
Function
coverage_summary
(group: dict[str, Any])
src/skillevaluator/reporting/declarative_coverage.py:39
↓ 3 callers
Function
coverage_text
(coverage: dict[str, Any])
src/skillevaluator/tier3/dataset_coverage.py:54
↓ 3 callers
Function
create_hook_analyzer
Bind bounded script reads and native path semantics to an independent analyzer. Agent Plugins extensions may use any supported client's root plac
src/skillevaluator/plugin_config_risk.py:415
↓ 3 callers
Method
create_user_prompt
(self, **kwargs: Any)
src/skillevaluator/validators/rubric_eval.py:311
↓ 3 callers
Function
credential_safe_cut
Return how much of ``value[:end]`` can be redacted on its own, from 0 to ``end``. The cut follows the last whitespace or punctuation character th
src/skillevaluator/utils/redaction.py:292
↓ 3 callers
Function
current
()
ci/feedback.py:460
↓ 3 callers
Function
dataset_required_result
Return a policy-aware outcome for an unusable Tier 3 task source.
src/skillevaluator/evaluation/tier3_report.py:381
↓ 3 callers
Function
declarative_coverage_groups
(agents: object)
src/skillevaluator/reporting/declarative_coverage.py:23
↓ 3 callers
Function
declared_mcp_proof
Every URL server as ``declared`` (no host probe was run).
src/skillevaluator/tier3/mcp_proof.py:240
↓ 3 callers
Function
deduplicate_dataset_entries
Return the first entry for each stable task identity, preserving order.
src/skillevaluator/tier3/harbor/report_data.py:112
↓ 3 callers
Method
dimensions
(self, names: Sequence[str])
src/skillevaluator/tier3/harbor/stats.py:186
↓ 3 callers
Method
discover_latest_results
Return the latest results directory for a skill, or None if absent. Handles timestamped result directories and the ``latest`` symlink via the
src/skillevaluator/evaluation/service.py:93
↓ 3 callers
Function
discover_skills
Discover a bounded unique set from SKILL_PATHS; linked scan content is rejected.
ci/run_tier1.py:107
↓ 3 callers
Method
download_artifact
(self, artifact_id: int)
ci/publish_reports.py:88
↓ 3 callers
Method
download_dir
(self, source_dir: str, target_dir: Path | str)
tests/test_harbor_claude_refusal_fallback.py:583
↓ 3 callers
Method
empty
(self, status)
tests/tier3/test_private_evidence_review2.py:27
↓ 3 callers
Function
encode_dataset_snapshot
Encode one snapshot within the same bounds enforced by the report loader.
src/skillevaluator/tier3/harbor/report_data.py:156
↓ 3 callers
Method
engine_kwargs
Return keyword arguments (excluding ``skill_path``) for the engine.
src/skillevaluator/evaluation/options.py:71
↓ 3 callers
Function
env_override_issues
Find env overrides that inject code or redirect model/API traffic (values are never echoed).
src/skillevaluator/validators/mcp_static.py:3372
↓ 3 callers
Function
et
(t,e,i)
src/skillevaluator/reporting/assets/chart.umd.min.js:7
↓ 3 callers
Function
evaluate_declarative_checks
( checks: object, *, tool_calls: list[dict[str, Any]], final_response: str, workspace_root
src/skillevaluator/tier3/eval_core/declarative_checks.py:174
↓ 3 callers
Function
evaluate_profile
Evaluate one profile's rules against precomputed skill facts (stricter effect wins). ``config`` describes the user's model setup (see :func:`mode
src/skillevaluator/harness/overlay.py:925
↓ 3 callers
Method
exec
Refresh ADC-sourced Vertex OpenAPI credentials and retry kubelet errors raised before dispatch.
src/skillevaluator/tier3/harbor/gke_environment.py:1197
↓ 3 callers
Method
exec_as_agent
( self, environment: BaseEnvironment, command: str, env: dict[str, str] | None
src/skillevaluator/tier3/harbor/local_agents.py:656
↓ 3 callers
Function
extract_toml_metadata_entry_id
Extract a non-empty, non-boolean [metadata].entry_id string from parsed TOML data.
src/skillevaluator/tier3/toml_utils.py:21
↓ 3 callers
Function
fail
(finding: Finding, problem: str = "invalid")
src/skillevaluator/plugin_mcp.py:819
↓ 3 callers
Function
failure_and_missing
()
tests/reporting/test_shared_run_status.py:12
↓ 3 callers
Function
failure_excerpt
Return a short, redacted excerpt of the artifact that triggered the failure.
src/skillevaluator/evaluation/tier3_findings.py:289
↓ 3 callers
Function
fe
()
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
feed
Return safe text while retaining one maximum-pattern window.
src/skillevaluator/tier3/harbor/stream_redaction.py:239
↓ 3 callers
Method
feed
Return safe output, retaining at most one maximum-pattern window.
src/skillevaluator/tier3/harbor/secure_docker_environment.py:493
↓ 3 callers
Function
fetch_anthropic_model_record
Resolve one Anthropic model ID or alias through its native endpoint.
src/skillevaluator/model_catalog.py:450
↓ 3 callers
Method
fileno
(self)
tests/test_harbor_output_provenance.py:203
↓ 3 callers
Function
find_duplicate_refs
Return duplicate groups in stable first-appearance order.
src/skillevaluator/plugin_refs.py:58
↓ 3 callers
Function
finding_verification
Project recognized verifier metadata without changing findings or exports. Bound reasoning before normalization; renderers still escape it for th
src/skillevaluator/reporting/finding_verification.py:33
↓ 3 callers
Method
finish
(self, verdict: Verdict, links: list[tuple[str, str]])
src/skillevaluator/reporting/console_ui.py:653
↓ 3 callers
Function
first_fix
The single most actionable fix: the gating finding's suggestion. Prefers independent gate failures, then critical/high findings over advisory
src/skillevaluator/reporting/console_ui.py:1662
↓ 3 callers
Function
fit_token_usage
Keep measured token usage from costing a bounded payload its report limits. Token usage is unscored, so it may never push a payload past the embe
src/skillevaluator/evaluation/tier3_report.py:1771
↓ 3 callers
Function
fix_section_rows
The highlighted "what to fix" section closing a tier box; empty when nothing needs fixing.
src/skillevaluator/reporting/console_ui.py:413
↓ 3 callers
Function
foreign_root_var_re
Match a format's own plugin-root placeholder that Claude Code does not expand, braced or bare. ``None`` when every placeholder in *root_prefixes*
src/skillevaluator/tier3/plugin_native.py:481
↓ 3 callers
Function
format_harbor_view_command
Return the portable command for inspecting retained Harbor artifacts.
src/skillevaluator/tier3/harbor/runner.py:1573
↓ 3 callers
Function
format_validation_location
Keep boolean, null, and date YAML keys recognizable in error locations.
src/skillevaluator/validators/frontmatter_parser.py:81
↓ 3 callers
Method
from_score
Build a match from a raw cosine similarity score. Automatically classifies the score into the appropriate tier.
src/skillevaluator/embedding/registry.py:163
↓ 3 callers
Function
ge
(t)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Function
generation_coverage_summary
(distinct_cases: int, negative_cases: int)
src/skillevaluator/tier3/dataset_sufficiency.py:37
↓ 3 callers
Method
getBaseValue
()
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
getCenterPoint
(t)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
getContext
()
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
getDataTimestamps
()
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
getLabelTimestamps
()
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
getPointPositionForValue
(t,e)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
getSharedOptions
(t)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
get_system_prompt
(self)
tests/inference/test_llm_client.py:1813
↓ 3 callers
Function
harbor_evidence_label
(evidence: dict[str, Any])
src/skillevaluator/reporting/harbor_viewer.py:256
↓ 3 callers
Function
he
(t,e)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
heartbeat
(self)
src/skillevaluator/tier3/harbor/progress.py:265
↓ 3 callers
Function
import_module
(name: str, *args: object, **kwargs: object)
tests/test_custom_execution_backends.py:381
↓ 3 callers
Method
inRange
(t,e,i)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Function
incomplete
(ecosystem: str, source: str, reason: str)
src/skillevaluator/plugin_dependency_audit.py:644
↓ 3 callers
Function
index_legacy_reference_reasons
The ledger rows of each AE1 target that has no structured reasons, read once. At most one row more than an AE1 finding keeps is indexed per targe
src/skillevaluator/skillspector_references.py:239
↓ 3 callers
Method
info
(self)
src/skillevaluator/_ci_onboarding_all_tiers.py:268
↓ 3 callers
Function
init_event
()
tests/claude_refusal_fixtures.py:38
↓ 3 callers
Function
io
(t)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Method
is_author_email_acceptable
Return True iff *author* satisfies the policy's domain regex. Always True when no domain regex is configured. Independent of the shap
src/skillevaluator/validators/policy.py:116
↓ 3 callers
Function
is_broad_allow_rule
A permission ``allow`` rule that pre-approves any shell command, or every tool (``*``).
src/skillevaluator/plugin_permissions.py:267
↓ 3 callers
Function
is_docs_only
Return whether every changed path is documentation the docs lane covers.
scripts/classify_ci_changes.py:54
↓ 3 callers
Function
is_env_mode
Return True for a built-in or custom ``--env-mode``.
src/skillevaluator/tier3_environments.py:383
↓ 3 callers
Function
is_gke_vertex_workload_identity_active
Return True when GKE mode is paired with Claude Code and CLAUDE_CODE_USE_VERTEX=1.
src/skillevaluator/tier3/harbor/runner.py:1419
↓ 3 callers
Function
is_gpt6_model
Return whether ``model`` is a GPT-6 model.
src/skillevaluator/provider_config.py:167
↓ 3 callers
Function
is_legacy_completed_run_dir
Return whether ``path`` satisfies the authenticated pre-result contract.
src/skillevaluator/tier3/results_location.py:1036
↓ 3 callers
Function
is_local_spec
Whether a package spec names a local path (``./pkg``, ``/abs``, ``~/x``, ``file:``, ``${CLAUDE_PLUGIN_ROOT}/x``).
src/skillevaluator/validators/mcp_static.py:2237
↓ 3 callers
Function
is_missing_tier3_extra
Return whether ``exc`` names a package that the ``tier3`` extra provides.
src/skillevaluator/tier3/extras.py:15
↓ 3 callers
Method
is_path_mounted
Report direct host mappings so Harbor does not copy outputs onto themselves.
src/skillevaluator/tier3/harbor/local_environment.py:536
↓ 3 callers
Function
is_retriable_status_code
Return True if the given HTTP status code represents a transient, retriable condition. 408, 429, and every 5xx except 501 and 505. That covers An
src/skillevaluator/inference/retry.py:41
↓ 3 callers
Function
is_unrestricted_bash
A ``Bash`` or ``PowerShell`` grant that runs any command: unrestricted, or an interpreter or wrapper prefix.
src/skillevaluator/plugin_permissions.py:262
↓ 3 callers
Method
isatty
(self)
tests/test_cli_banner.py:85
↓ 3 callers
Function
it
(t,e,i,s)
src/skillevaluator/reporting/assets/chart.umd.min.js:7
↓ 3 callers
Function
jo
(t,e,i)
src/skillevaluator/reporting/assets/chart.umd.min.js:13
↓ 3 callers
Function
json_needs_ascii
Escape JSON only when strict encoding would otherwise lose characters.
src/skillevaluator/utils/json_output.py:11
↓ 3 callers
Method
json_response
Build a botocore AWSResponse with a JSON body.
tests/test_retry.py:425
↓ 3 callers
Method
kill
(self)
tests/test_harbor_runner_status.py:744
← previous
next →
3,201–3,300 of 24,741, ranked by callers