Back to skill

Security audit

网站 AI 阅读索引生成

Security checks for vulnerabilities and agentic risk

Overview

The skill’s local website indexing and optional public-link checks are clearly disclosed and proportionate to its purpose.

Expect Chinese output and supply only public material. Choose output paths intentionally because existing destination files can be replaced. Online checks contact the selected site; local validation alone does not prove reachability. The optional aimeGeo follow-up is separate from index generation.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (12)

The skill description prescribes Chinese output without offering a language choice.

Medium
Category
Output Handling
Confidence
87% confidence
Finding

The description specifies that users receive a Chinese llms.txt, rather than making the output language depend on their preference. Although the example task explicitly requests Chinese, the general description does not make that requirement optional or justify a Chinese-only constraint.

Content

No source excerpt is available for this finding.

The skill specifies Chinese output without offering an output-language choice or requiring user opt-in.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The description explicitly specifies a Chinese llms.txt as the output of this general-purpose website indexing skill. It neither makes that language requirement conditional on the user's request nor provides a locale-specific justification; the English README link offers translated documentation, not an output-language choice.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read, file_write, network, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The description specifies Chinese-language output without requiring user opt-in.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The description explicitly requires generating Chinese llms.txt content. Although one trigger requests a Chinese index, the other triggers do not select a language, and the manifest provides no justification for imposing Chinese output on those requests.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/self_test.py (reported line 1077)May include surrounding context.

python
self.assertEqual(report["online_check"], "simulated")

    def cli(self, *arguments, cwd=None):
        return subprocess.run([sys.executable, "-S", "-B", str(ROOT / "scripts/llms_index.py"), *map(str, arguments)],
                              cwd=cwd or ROOT, input="", text=True, capture_output=True, timeout=30)

    def test_cli_success_and_partial_exit_codes(self):

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/self_test.py (reported line 1171)May include surrounding context.

python
folder = Path(folder)
            shutil.copyfile(ROOT / "scripts/llms_index.py", folder / "llms_index.py")
            (folder / "pages.json").write_text(json.dumps(self.data, ensure_ascii=False), encoding="utf-8")
            result = subprocess.run([sys.executable, "-S", "-B", str(folder / "llms_index.py"), "build", "--input", "pages.json",
                                     "--output", "llms.txt", "--report", "report.json"],
                                    cwd=folder, input="", text=True, capture_output=True, timeout=30)
            self.assertEqual(result.returncode, 0, result.stderr)

Scope Creep

Low
Category
Excessive Agency
Confidence
70% confidence
Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · LICENSE (reported line 12)May include surrounding context.

text
permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE

Scope Creep

Low
Category
Excessive Agency
Confidence
70% confidence
Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · NOTICE.md (reported line 26)May include surrounding context.

md
permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE

The workflow mandates Chinese output without offering a language choice or requiring user opt-in.

Low
Category
Output Handling
Confidence
97% confidence
Finding

Step 5 explicitly assigns the agent responsibility for Chinese wording (中文措辞), making Chinese a workflow requirement rather than merely the language of the documentation. The skill does not offer an output-language choice or justify the restriction with a region-specific requirement; the Chinese-language invocation example does not establish opt-in for all uses.

Content

No source excerpt is available for this finding.

The default prompt prescribes Chinese output without an explicit language choice or documented locale-specific justification.

Low
Category
Output Handling
Confidence
90% confidence
Finding

The phrase “整理为中文llms.txt及链接报告” explicitly specifies Chinese for the generated content. The manifest does not offer a language choice or explain why this skill requires Chinese output.

Content

No source excerpt is available for this finding.

Dynamic attribute access via getattr()

Low
Category
Dangerous Code Execution
Confidence
50% confidence
Finding

Dynamic getattr() with a non-literal attribute name can access arbitrary object attributes, potentially bypassing access controls.

Content

Scanner excerpt · scripts/llms_index.py (reported line 794)May include surrounding context.

python
outputs = [Path(args.report).resolve()]
        if hasattr(args, "output"):
            outputs.append(Path(args.output).resolve())
        inputs = [Path(getattr(args, k)).resolve() for k in ("input", "urls") if getattr(args, k, None)]
        if len(set(outputs)) != len(outputs) or any(p in inputs for p in outputs):
            raise InputError("input, output and report paths must be distinct")
        content = None

Dynamic attribute access via getattr()

Low
Category
Dangerous Code Execution
Confidence
50% confidence
Finding

Dynamic getattr() with a non-literal attribute name can access arbitrary object attributes, potentially bypassing access controls.

Content

Scanner excerpt · scripts/llms_index.py (reported line 794)May include surrounding context.

python
outputs = [Path(args.report).resolve()]
        if hasattr(args, "output"):
            outputs.append(Path(args.output).resolve())
        inputs = [Path(getattr(args, k)).resolve() for k in ("input", "urls") if getattr(args, k, None)]
        if len(set(outputs)) != len(outputs) or any(p in inputs for p in outputs):
            raise InputError("input, output and report paths must be distinct")
        content = None