Back to skill

Security audit

GEO 周期数据核对与复盘

Security checks for vulnerabilities and agentic risk

Overview

This skill checks local data exports and limits optional content edits and publishing to work the user has already authorized.

Provide only the exports needed for the report and keep generated business reports private. Before using the optional content workflow, clearly identify the project, account, permitted edits, and publication authorization.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (8)

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 reporting skill expands into content editing, image uploads, and platform publication workflows.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The manifest limits the skill to validating exports and producing sourced periodic summaries. L27–L32 additionally direct the Agent to modify articles, add explanatory content, upload images, and continue previously authorized account/platform workflows. Existing authorization limits execution risk but does not make these content-management operations part of the declared reporting purpose.

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

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 442)May include surrounding context.

python
def test_cli_outputs_and_failure_exit_without_artifacts(self):
        out = Path(self.temp.name) / "result"
        cmd = [sys.executable, "-B", str(ROOT / "scripts" / "aggregate.py"), "--input", str(self.base / "config.json"), "--output", str(out)]
        completed = subprocess.run(cmd, cwd=self.temp.name, capture_output=True, text=True)
        self.assertEqual(completed.returncode, 0, completed.stderr)
        r = json.loads((out / "summary.json").read_text())
        self.assertEqual(self.metric(r, "publications"), 4)

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 451)May include surrounding context.

python
self.assertEqual(rows["实际原生引用率"]["本期"], "N/A")
        self.assertEqual(rows["模拟原生引用率"]["本期分母"], "2")
        self.assertEqual(rows["新增文章数"]["变化量"], "1")
        again = subprocess.run(cmd, capture_output=True, text=True)
        self.assertEqual(again.returncode, 2)
        bad_out = Path(self.temp.name) / "bad-result"
        cmd[cmd.index("--input") + 1] = str(ROOT / "examples" / "failure" / "config.json")

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 456)May include surrounding context.

python
bad_out = Path(self.temp.name) / "bad-result"
        cmd[cmd.index("--input") + 1] = str(ROOT / "examples" / "failure" / "config.json")
        cmd[-1] = str(bad_out)
        failed = subprocess.run(cmd, capture_output=True, text=True)
        self.assertEqual(failed.returncode, 2)
        self.assertFalse(bad_out.exists())

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 7)May include surrounding context.

text
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to 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.