Back to skill

Security audit

FlowUs 文件版本管理器

Security checks for vulnerabilities and agentic risk

Overview

ClawScan’s verdict is benign with medium confidence: the scanners found no AIG issues and judged the skill’s purpose and persistence behavior acceptable, while noting permission scope, installation considerations, and incomplete inspection.

The verdict is benign, but the evidence is incomplete: Skillspector inspected 27 of 28 components and partially interpreted one, while its tool misuse analyzer was degraded. Before installing, consider whether the skill’s detected shell, file, network, and environment capabilities fit your needs, and review its connection behavior and authentication settings.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (9)

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

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/flowus_version_manager/cli_transport.py (reported line 30)May include surrounding context.

python
def _run(self, arguments, *, raw=False):
        try:
            result = subprocess.run([self.executable, "--base-url", "https://api.flowus.cn", "--json", "--no-retry", *arguments],
                                    capture_output=True, text=True, timeout=60, check=False)
            if len(result.stdout) > 16 * 1024 * 1024:
                raise ManagerError("FlowUs CLI response exceeds its safety bound")

Unsafe Defaults

Medium
Category
Tool Misuse
Confidence
75% confidence
Finding

Tool defaults are unsafe or overly permissive (e.g. disabled TLS verification, no authentication, world-writable permissions). Unsafe defaults widen the attack surface.

Content

Scanner excerpt · scripts/flowus_version_manager/profiles.py (reported line 251)May include surrounding context.

python
return result


def connect(path, root, transport, *, agent_id=None, writer=None, dry_run=False, replace=False, existing=None, rest_auth=None):
    root = root_id(root)
    if transport not in {"rest", "mcp_relay"}:
        raise ManagerError("Choose an explicit connection transport")

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/flowus_version_manager/rules.py (reported line 46)May include surrounding context.

python
tool = shutil.which("pdftotext" if suffix == ".pdf" else "textutil")
            if tool and suffix != ".xls":
                command = [tool, str(path), "-"] if suffix == ".pdf" else [tool, "-convert", "txt", "-stdout", str(path)]
                result = subprocess.run(command, capture_output=True, timeout=20, check=False)
                if result.returncode == 0:
                    return result.stdout.decode("utf-8", errors="replace"), warnings
            binary = shutil.which("strings") if suffix in {".xls", ".doc"} else None

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/flowus_version_manager/rules.py (reported line 51)May include surrounding context.

python
return result.stdout.decode("utf-8", errors="replace"), warnings
            binary = shutil.which("strings") if suffix in {".xls", ".doc"} else None
            if binary:
                result = subprocess.run([binary, "-n", "3", str(path)], capture_output=True, timeout=20, check=False)
                if result.returncode == 0:
                    return result.stdout.decode("utf-8", errors="replace"), ["Legacy binary text extraction is approximate"]
            return "", ["Legacy/PDF text extraction unavailable; file retained and metadata matching only"]

Dynamic import via __import__()

Medium
Category
Dangerous Code Execution
Confidence
75% confidence
Finding

Dynamic import() can load arbitrary modules at runtime, bypassing static analysis and potentially importing malicious code.

Content

Scanner excerpt · scripts/flowus_version_manager/store.py (reported line 83)May include surrounding context.

python
def decode(self, page, headers=None):
        props = page.get("properties", {})
        raw = plain(props.get("Payload", {}).get("rich_text"))
        data = __import__("json").loads(raw) if raw else {}
        if not isinstance(data, dict):
            raise ManagerError("Cloud Payload must be a JSON object")
        for field, name in [("Name", "title"), ("Status", "status"), ("Category", "category"),

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 · scripts/flowus_version_manager/transport.py (reported line 14)May include surrounding context.

python
from .common import APIError, Conflict, ManagerError, NetworkError, canonical
from . import __version__

API_BASE = "https://api.flowus.cn/v2"
MAX_FILE = 100 * 1024 * 1024

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/flowus_version_manager/cli.py (reported line 275)May include surrounding context.

python
"status": "partial_failure" if summary["failed"] else "completed"}
        return result if args.dry_run or not any(summary[k] for k in ('added', 'duplicates', 'needs_review')) else after_write(manager, result)
    if args.command in ("latest", "timeline"):
        return getattr(manager, args.command)(args.query)
    if args.command == "ask":
        return manager.ask(args.question)
    if args.command == "note":

The question-answering response is hard-coded in Chinese without a language choice.

Low
Category
Output Handling
Confidence
92% confidence
Finding

Both possible values of the user-facing answer field are Chinese, regardless of the question's language or the user's preferences. The file provides no language opt-in or documented justification for this fixed-language output.

Content

No source excerpt is available for this finding.