NVIDIA/SkillSpector

Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, security risks, prompt injection, data exfiltration, and supply-chain risks in Claude Code, Codex, and MCP skills before you install them.

Python★ 22,395+214 that day#4 on trending

Analyzed October 27, 2026 at d7338f5 (main) with jscpd v5 in 1.9s.

E32.2/100

Project health209,018 lines of code (L)

Duplication69

4.7% in python, typescript, javascript, makefile, bash, docker

Dead code15

20.5%

Complexity32

82.8%

The score is jscpd's own --health model: markup, data and text left out, tests kept. The numbers below are the trending scan's.

2.91%duplicated code · 2,539 of 87,188 lines

172clones in code

115code files of 446 scanned

4.46%duplicated over all files · 779 clones

Most complex files

Dead code 20.48% unused · 49 findings

src/skillspector/nodes/analyzers/static_patterns_tool_misuse.pyunused-file
src/skillspector/nodes/analyzers/static_patterns_supply_chain.pyunused-file
src/skillspector/nodes/analyzers/static_runner.pyunused-file
src/skillspector/nodes/analyzers/behavioral_taint_tracking.pyunused-file
src/skillspector/nodes/analyzers/static_python_shell_truthiness.pyunused-file
src/skillspector/nodes/analyzers/bundled_execution_surface.pyunused-file
src/skillspector/security_reconstruction.pyunused-file
src/skillspector/nodes/analyzers/mcp_tool_poisoning.pyunused-file
src/skillspector/nodes/analyzers/static_yara.pyunused-file
src/skillspector/nodes/analyzers/static_patterns_privilege_escalation.pyunused-file

The code number leaves out tests, fixtures, docs, examples, data files, styles, translations and vendored code, and 1 generated file: src/skillspector/suppression.py.

Largest clones in the code

169 lines in 2 placespython790 tokens

src/skillspector/nodes/analyzers/behavioral_ast.py:653–821src/skillspector/nodes/analyzers/behavioral_taint_tracking.py:2284–2452

653"""Parse Python files via AST and detect dangerous execution patterns."""
654components: list[str] = state.get("components") or []
655file_cache: dict[str, str] = state.get("local_file_cache") or state.get("file_cache") or {}
656raw_file_cache = state.get("raw_file_cache")
657source_classifications = (
658    state.get("python_source_classifications")
659    if "python_source_classifications" in state
660    else None
661)
662source_classification_limitations = state.get("python_source_classification_limitations") or {}

First 10 of 169 lines.

65 lines in 2 placespython354 tokens

src/skillspector/nodes/analyzers/semantic_developer_intent.py:251–315src/skillspector/nodes/analyzers/semantic_quality_policy.py:238–302

251    base_prompt=prompt,
252    model=model,
253    node=ANALYZER_ID,
254    timeout=timeout,
255)
256batches = analyzer.get_batches(files, file_cache)
257results = run_async(analyzer.arun_batches(batches))
258outcome = getattr(analyzer, "_last_batch_outcome", BatchExecutionResult(successful=results))
259findings = analyzer.collect_findings(outcome.successful)
260events, status = ledger_events_for_batches(ANALYZER_ID, outcome)

First 10 of 65 lines.

11 lines in 6 placespython62 tokens
src/skillspector/nodes/analyzers/static_patterns_data_exfiltration.py:514–524 src/skillspector/nodes/analyzers/static_patterns_excessive_agency.py:434–444 src/skillspector/nodes/analyzers/static_patterns_memory_poisoning.py:631–641 src/skillspector/nodes/analyzers/static_patterns_memory_poisoning.py:658–668 src/skillspector/nodes/analyzers/static_patterns_output_handling.py:679–689and 1 more
```
514        message="File System Enumeration",
515        severity=Severity.MEDIUM,
516        location=loc(line_num),
517        confidence=confidence,
518        tags=tag,
519        context=ctx(match.start()),
520        matched_text=match.group(0)[:200],
521        complete_match=match.group(0),
522    )
523)
```
First 10 of 11 lines.

7 more

10 lines in 6 placespython57 tokens
src/skillspector/nodes/analyzers/static_patterns_agent_snooping.py:164–173 src/skillspector/nodes/analyzers/static_patterns_data_exfiltration.py:532–541 src/skillspector/nodes/analyzers/static_patterns_supply_chain.py:2312–2321 src/skillspector/nodes/analyzers/static_patterns_supply_chain.py:2329–2338 src/skillspector/nodes/analyzers/static_patterns_system_prompt_leakage.py:984–993and 1 more
10 lines in 6 placespython57 tokens
src/skillspector/nodes/analyzers/static_patterns_data_exfiltration.py:514–523 src/skillspector/nodes/analyzers/static_patterns_data_exfiltration.py:549–558 src/skillspector/nodes/analyzers/static_patterns_excessive_agency.py:505–514 src/skillspector/nodes/analyzers/static_patterns_output_handling.py:700–709 src/skillspector/nodes/analyzers/static_patterns_rogue_agent.py:356–365and 1 more
52 lines in 2 placespython253 tokens
src/skillspector/nodes/analyzers/behavioral_ast.py:240–291 src/skillspector/nodes/analyzers/behavioral_taint_tracking.py:228–279
37 lines in 2 placespython233 tokens
src/skillspector/nodes/report.py:691–727 src/skillspector/nodes/report.py:737–773
10 lines in 5 placespython55 tokens
src/skillspector/nodes/analyzers/static_patterns_agent_snooping.py:142–151 src/skillspector/nodes/analyzers/static_patterns_data_exfiltration.py:432–441 src/skillspector/nodes/analyzers/static_patterns_excessive_agency.py:412–421 src/skillspector/nodes/analyzers/static_patterns_memory_poisoning.py:612–621 src/skillspector/nodes/analyzers/static_patterns_output_handling.py:622–631
26 lines in 2 placespython205 tokens
src/skillspector/nodes/analyzers/static_patterns_memory_poisoning.py:362–387 src/skillspector/nodes/analyzers/static_patterns_system_prompt_leakage.py:874–899
44 lines in 2 placespython189 tokens
src/skillspector/nodes/analyzers/behavioral_ast.py:610–653 src/skillspector/nodes/analyzers/behavioral_taint_tracking.py:2241–2284

Largest clones over all files, tests and data included

169 lines in 2 placespython790 tokens

src/skillspector/nodes/analyzers/behavioral_ast.py:653–821src/skillspector/nodes/analyzers/behavioral_taint_tracking.py:2284–2452

653"""Parse Python files via AST and detect dangerous execution patterns."""
654components: list[str] = state.get("components") or []
655file_cache: dict[str, str] = state.get("local_file_cache") or state.get("file_cache") or {}
656raw_file_cache = state.get("raw_file_cache")
657source_classifications = (
658    state.get("python_source_classifications")
659    if "python_source_classifications" in state
660    else None
661)
662source_classification_limitations = state.get("python_source_classification_limitations") or {}

First 10 of 169 lines.

7 lines in 11 placespythontests52 tokens

tests/nodes/analyzers/test_security_reconstruction.py:537–543tests/nodes/analyzers/test_security_reconstruction.py:2034–2040tests/nodes/analyzers/test_security_reconstruction.py:2054–2060tests/nodes/analyzers/test_security_reconstruction.py:2124–2130tests/nodes/analyzers/test_security_reconstruction.py:2310–2316and 6 more

537result = static_runner.run_static_patterns_with_ledger(
538    {"components": ["SKILL.md"], "file_cache": {"SKILL.md": content}}, [tm_module]
539)
540
541event = result["inspection_ledger"][0]
542assert event["outcome"] is LedgerOutcome.PARTIAL
543assert event["reason_code"] is LedgerReason.OBFUSCATED_INSTRUCTION_TEXT

2034result = static_runner.run_static_patterns_with_ledger(
2035    {"components": ["SKILL.md"], "file_cache": {"SKILL.md": content}}, [tm_module]
2036)
2037
2038event = result["inspection_ledger"][0]
2039assert event["outcome"] is LedgerOutcome.PARTIAL
2040assert event["reason_code"] is LedgerReason.STATIC_PARSE_LIMIT

12 lines in 8 placespythontests57 tokens
tests/nodes/test_semantic_quality_policy.py:819–830 tests/nodes/test_semantic_quality_policy.py:851–862 tests/nodes/test_semantic_quality_policy.py:879–890 tests/nodes/test_semantic_quality_policy.py:937–948 tests/nodes/test_semantic_quality_policy.py:965–976and 3 more
```
819    pytest.skip("sqp1_vague_triggers fixture not present")
820
821file_cache = _build_file_cache(skill_dir)
822state: dict = {"file_cache": file_cache}
823
824from skillspector.llm_analyzer_base import LLMAnalyzerBase
825
826orig_init = LLMAnalyzerBase.__init__
827
828def _patched_init(self_inner, *args, **kwargs):
```
First 10 of 12 lines.
9 lines in 8 placespythontests55 tokens
tests/test_mcp_tool_poisoning.py:372–380 tests/test_mcp_tool_poisoning.py:390–398 tests/test_mcp_tool_poisoning.py:405–413 tests/test_mcp_tool_poisoning.py:426–434 tests/test_mcp_tool_poisoning.py:445–453and 3 more
65 lines in 2 placespython354 tokens
src/skillspector/nodes/analyzers/semantic_developer_intent.py:251–315 src/skillspector/nodes/analyzers/semantic_quality_policy.py:238–302

By format, code only

Format	Files	Lines	Clones	Duplicated lines	Duplication
python	112	86,259	172	2,539	2.94%
typescript	3	929	0	0	0%

By format, all files

Format	Files	Lines	Clones	Duplicated lines	Duplication
python	313	206,722	766	9,665	4.68%
typescript	4	1,794	6	71	3.96%
markdown	74	9,586	3	37	0.39%
yaml	18	1,139	3	90	7.9%
json	5	159	1	14	8.81%
text	16	567	0	0	0%
javascript	1	432	0	0	0%
bash	9	380	0	0	0%
mermaid	3	178	0	0	0%
makefile	1	158	0	0	0%

Trending appearances

Day	Rank	Stars	Files	Lines	Clones	All files	Code	Health	Commit
Oct 27, 2026	#4	22,395 (+214)	446	221,258	779	4.46%	2.91%	E 32.2	`d7338f5`

Want this for your own project? Install jscpd and run jscpd ., or measure it the way the baselines were measured.