Back to skill

Security audit

递延所得税·逐条挂源

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious with medium confidence because scanners report behavior outside its deferred-tax document purpose, including external parsing and persistent home-directory usage markers.

Consider the reported external document parsing and tool-discovery activity, and the automatic local usage markers, before installing. The skillspector scan was partial: it inspected all eight listed components but left one local reference unresolved, so the evidence is not complete. The separate AIG scan completed with no findings; that does not establish that the skill is safe.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
Findings (17)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · README.md (reported line 149)May include surrounding context.

md
**它不会碍事:** 任何失败(家目录只读、配额满、无 HOME)都静默跳过,不报错、不改退码、不吐一句杂音。同一天跑多次只留一枚。

要清掉:`rm -rf ~/.cue-usage`。

---

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · README.md (reported line 149)May include surrounding context.

md
**它不会碍事:** 任何失败(家目录只读、配额满、无 HOME)都静默跳过,不报错、不改退码、不吐一句杂音。同一天跑多次只留一枚。

要清掉:`rm -rf ~/.cue-usage`。

---

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The reported automatic usage marker is unrelated to document validation and persists activity information outside the document workspace. This source mentions beacon testing but does not explain the home-directory write or provide a consent or disable mechanism; an empty dated marker creates only limited local privacy exposure, not demonstrated exfiltration.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared purpose is source-grounded deferred-tax disclosure extraction, while the supplied code performs service and tool discovery. No financial-report input, disclosure parsing, quotation anchoring, or register generation is present, nor is there evidence connecting the discovered tools to those tasks. On the supplied evidence, these network operations are not established as supporting implementation details. ROUTING_URL is not defined in the chunk, so successful execution of the second request cannot be assumed.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

Persistent, date-deduplicated usage files introduce an unnecessary local activity record beyond the financial-report task. The source does not meaningfully disclose or gate that behavior, although the limited marker contents do not justify the static finding's high severity.

Content

No source excerpt is available for this finding.

Env Variable Harvesting

High
Category
Data Exfiltration
Confidence
60% confidence
Finding

Code enumerates, copies, or searches environment variables for secrets. Bulk environment access can collect credentials unrelated to the skill's stated purpose.

Content

Scanner excerpt · tests/test_deferred_tax.py (reported line 31)May include surrounding context.

python
def run(*args, env=None, cwd=None):
    e = dict(os.environ, PYTHONDONTWRITEBYTECODE="1")
    if env:
        e.update(env)
    return subprocess.run([sys.executable, "-B", str(args[0]), *map(str, args[1:])],

Env Variable Harvesting

High
Category
Data Exfiltration
Confidence
60% confidence
Finding

Code enumerates, copies, or searches environment variables for secrets. Bulk environment access can collect credentials unrelated to the skill's stated purpose.

Content

Scanner excerpt · tests/test_deferred_tax.py (reported line 179)May include surrounding context.

python
os.chmod(td, 0o700)

    def test_no_home_is_silent(self):
        e = dict(os.environ, PYTHONDONTWRITEBYTECODE="1")
        e.pop("HOME", None)
        r = subprocess.run([sys.executable, "-B", str(CHECK), "--demo"],
                           capture_output=True, text=True, env=e, cwd=str(PKG))

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · README.md (reported line 53)May include surrounding context.

CATALOG="https://cuecue.cn/api/mcp-catalog" curl -s "$CATALOG" | head -c 2000

从目录里取本组的路由地址,确认工具在场

curl -s -X POST "$ROUTING_URL"
-H 'Content-Type: application/json'
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

text

The attachment-parsing workflow lacks a warning that supplied document contents will be sent to an external service.

Medium
Category
Output Handling
Confidence
89% confidence
Finding

The workflow directs attachments and scans to an external parsing channel without explicitly warning that their contents may leave the local environment or requiring approval. Although published filings are usually public, supported user-supplied documents could contain confidential financial information.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, file_read, file_write, 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 workflow sends attachments to an external parsing service without warning that user-provided document contents may leave the local environment.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The attachment workflow initiates external parsing when evidence resides in an attachment, without an explicit document-transfer warning or approval requirement. Public annual reports lower the typical sensitivity, but user-supplied PDFs and screenshots may contain nonpublic financial data.

Content

No source excerpt is available for this finding.

L40 的登记稿作出递延所得税资产可转回及税务风险可控的判断,超出清单登记范围。

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

Manifest 明确排除判定递延所得税资产能否转回及评价税务风险,但此行直接宣称“递延所得税资产能够转回,税务风险可控”。所附原句仅披露本期所得税费用,不能支持这两项判断;开头的人工核对提示也不能消除这一范围冲突。

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 · tests/test_deferred_tax.py (reported line 34)May include surrounding context.

python
e = dict(os.environ, PYTHONDONTWRITEBYTECODE="1")
    if env:
        e.update(env)
    return subprocess.run([sys.executable, "-B", str(args[0]), *map(str, args[1:])],
                          capture_output=True, text=True, env=e, cwd=str(cwd or PKG))

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 · tests/test_deferred_tax.py (reported line 181)May include surrounding context.

python
def test_no_home_is_silent(self):
        e = dict(os.environ, PYTHONDONTWRITEBYTECODE="1")
        e.pop("HOME", None)
        r = subprocess.run([sys.executable, "-B", str(CHECK), "--demo"],
                           capture_output=True, text=True, env=e, cwd=str(PKG))
        self.assertEqual(r.returncode, 0)
        self.assertEqual(r.stderr, "", f"无 HOME 不许吐杂音:{r.stderr[:200]}")

Undeclared usage tracking writes persistent files outside the document workflow.

Low
Category
Excessive Agency
Confidence
99% confidence
Finding

mark_first_run automatically creates a dated skill-usage marker in the home directory before validation, demo output, or self-testing, with no opt-out. The function documents its purpose internally and performs no network transmission, but it still creates an unnecessary persistent local activity record outside the requested document workflow.

Content

No source excerpt is available for this finding.

Home-directory resolution contradicts the explicit claim that no environment variables are read.

Low
Category
Excessive Agency
Confidence
99% confidence
Finding

The module documentation at L003 explicitly promises 不读环境变量 (no environment-variable reads). os.path.expanduser("~") normally reads HOME on Unix or environment-based profile information on Windows, so the tracking implementation violates that promise even though it does not access os.environ directly.

Content

No source excerpt is available for this finding.

The tests require a home-directory usage beacon unrelated to deferred-tax extraction or validation.

Low
Category
Excessive Agency
Confidence
95% confidence
Finding

These tests require the checker to persist a skill-usage marker in the user’s home directory, including during demo execution, although tracking is unrelated to deferred-tax validation. That introduces unnecessary filesystem writes and a local usage trace; the shown tests establish only an empty local marker, not network transmission or credential collection.

Content

No source excerpt is available for this finding.