Back to skill

Security audit

Thermal Relative Fever Screening (Multi-Person Gathering) | 家庭多人聚集时体温相对异常检测

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because scanners report a mismatch with its thermal-screening purpose, undeclared identity and account handling, and behavior that may trigger remote processing unexpectedly.

Consider avoiding installation until the skill’s maintainer explains the account and credential handling, remote destinations, local video uploads, and the behavior of `--export-env-only`. The scan is incomplete because one local reference was missing; the available results do not establish malicious intent, but they also do not resolve the reported concerns.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
Findings (25)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The supplied code's substantive purpose is user-account database management rather than thermal analysis. Database utilities could support a larger thermal-monitoring application, but this chunk specifically provides identity and token persistence with no demonstrated connection to temperature measurements or alerts. That makes the difference more substantial than an ordinary supporting implementation detail. The assessment is limited to this chunk; it does not establish that thermal functionality is absent elsewhere. No network transmission or token exfiltration is shown.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

Generic formatting, polling, and filesystem helpers could reasonably support the declared feature and are not independently grounds for a mismatch. However, this chunk's substantive account provisioning, credential handling, authenticated networking, and billing behavior goes beyond the stated camera-temperature analysis purpose. None of the declared thermal-analysis behavior is implemented or explicitly invoked here. This finding is limited to the supplied chunk; other files could implement the advertised feature.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The visible behavior does not implement the declared purpose. The agent invocation is disabled, so the chunk does not actually launch an external agent. Imported base classes and services are not supplied, so their behavior cannot be verified; this assessment is limited to the shown code.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied code implements a generic video-analysis and report-management client rather than the declared thermal-screening workflow. It contains no visible camera integration or temperature-specific processing. The imported analysis service could implement additional functionality, but that implementation is not supplied, so the code does not establish the claimed temperature-analysis behavior. File/URL submission and report-history access are concrete capabilities absent from the description.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding

The supplied chunk implements a generic API client rather than thermal-image processing, temperature comparison, or alert generation. Network requests could support the declared system, but this chunk exposes additional capabilities—especially user lookup and object download-URL retrieval—with no visible connection to that purpose. This assessment is limited to the supplied code: it does not establish malicious activity, actual endpoint destinations, or whether thermal-analysis functionality exists elsewhere.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding

The supplied code contains no visible thermal-camera integration, per-person temperature extraction, comparison against other people's temperatures, threshold evaluation, or temperature-specific warning. Analysis is delegated to an imported implementation that was not supplied, so its internal behavior cannot be verified. Independently, the explicit --list operation retrieves user-associated analysis history, an additional capability outside the declared temperature-monitoring purpose. Routine output-file writing is a supporting implementation detail and is not treated as a separate mismatch.

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, 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 --export-env-only option promises not to execute analysis, but the code ignores it and runs the analysis pipeline.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The parser accepts an option explicitly promising no analysis, but the program never checks it before identity initialization or the analysis/list branches. With an input supplied, a user expecting configuration-only behavior can unintentionally initiate remote processing of sensitive household or health-related video.

Content

No source excerpt is available for this finding.

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · skills/smyx_analysis/scripts/skill.py (reported line 28)May include surrounding context.

python
result_json = JsonUtil.parse(result_json_pure_text, result_json_pure_text)

        result_json_common_ai_response = result_json.get("commonAiResponse") if isinstance(result_json,
                                                                                           dict) else result_json
        if result_json_common_ai_response:
            result_json = result_json_common_ai_response

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · skills/smyx_analysis/scripts/skill.py (reported line 33)May include surrounding context.

python
result_json = result_json_common_ai_response

        result_json_health_ai_response = result_json.get("healthAiResponse") if isinstance(result_json,
                                                                                           dict) else result_json
        if result_json_health_ai_response:
            result_json = result_json_health_ai_response

Local video contents are passed to the analysis API without disclosure that the file will be uploaded.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The local-file branch reads the complete video and supplies it as a multipart file to the analysis service without an upload notice or consent check at this boundary. Although API use is documented, that does not clearly establish informed authorization to transfer household footage and health-related information to a remote processor.

Content

No source excerpt is available for this finding.

The --export-env-only option promises not to analyze video, but the code ignores it and proceeds with analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The --export-env-only flag is parsed but never used, so supplying it with a video input still invokes analysis. This violates an explicit no-analysis expectation and can cause unintended remote processing, privacy exposure, or service charges.

Content

No source excerpt is available for this finding.

The service exposes username-based user lookup, which is unrelated to the declared thermal-screening purpose.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

At L098-L101, get_user_by_username delegates a username-based user lookup to RequestUtil. The manifest describes comparing temperatures among people in a scene and issuing relative-temperature alerts; it does not describe account lookup or identity-based processing.

Content

No source excerpt is available for this finding.

The skill includes identity and credential database management unrelated to thermal screening.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The User model at L448–L465 stores names, email addresses, demographic information, and token/open_token credentials, while UserDao at L470–L473 exposes inherited database read, update, and delete operations over those records. Managing persistent user identities and credentials is not justified by the manifest's scene-based temperature comparison and retesting alerts, which do not require identifying people or managing their accounts.

Content

No source excerpt is available for this finding.

FileUtil.open silently opens files in write mode, truncating any existing contents.

Medium
Category
Output Handling
Confidence
97% confidence
Finding

Opening a path with mode 'w' immediately truncates an existing file. This generically named helper provides no comment, docstring, confirmation, or user-facing message explaining that destructive behavior; exception reporting does not disclose successful truncation.

Content

No source excerpt is available for this finding.

Thermal-screening requests can silently register remote user accounts and persist account credentials.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

When tokens are missing, an ordinary HTTP request can trigger remote login or registration with silent=1 and register=1, then save returned credentials through UserDao. This introduces undisclosed account creation and credential retention into a thermal-screening workflow whose description does not disclose an identity-enrollment requirement.

Content

No source excerpt is available for this finding.

The HTTP helper silently performs an additional identity-transmitting login or registration request.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The general-purpose HTTP helper conditionally sends an additional request to a health-service login endpoint, transmitting the selected identity as both openId and mobile. This hidden side effect can disclose a configured identifier and create a remote account without a distinct enrollment decision, which is particularly unexpected for a thermal-analysis skill.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
92% confidence
Finding

The POST transmits the selected username or identity value and skill-source metadata to a configured external login service without a separate consent gate. External requests are not inherently unsafe, but this one implements the undisclosed identity-registration behavior; the supplied source does not establish that the destination is malicious or that thermal imagery is transmitted here.

Content

Scanner excerpt · skills/smyx_common/scripts/util.py (reported line 561)May include surrounding context.

python
"source": ConstantEnum.DEFAULT__SKILL_HUB_NAME
            }
            try:
                _response = requests.post(_url, json=_data)
                if _response.status_code == 200:
                    _response_json = _response.json()
                    if _response_json and _response_json.get("success"):

Loading configuration silently creates directories and files, including during module import.

Low
Category
Output Handling
Confidence
90% confidence
Finding

YamlUtil.load creates parent directories and writes a YAML file when the requested configuration is absent, without a warning, log, docstring, or explanatory comment. Because BaseEnum.init_subclass calls this loader during class definition (L052–L059), merely importing the module can modify the filesystem, an unexpected side effect for a configuration read.

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 · skills/smyx_common/scripts/dao.py (reported line 243)May include surrounding context.

python
username = getattr(model, "username", None)
            if username:
                column_names = self.__model__.__table__.columns.keys()
                update_data = {key: getattr(model, key) for key in column_names if key != "username"}
                return self.update_by_username(username, **update_data)

            return 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 · skills/smyx_common/scripts/dao.py (reported line 326)May include surrounding context.

python
if filters:
                for key, value in filters.items():
                    query = query.filter(getattr(self.__model__, key) == value)

            if offset:
                query = query.offset(offset)

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 · skills/smyx_common/scripts/dao.py (reported line 441)May include surrounding context.

python
if filters:
                for key, value in filters.items():
                    query = query.filter(getattr(self.__model__, key) == value)

            return query.scalar()
        finally:

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 · skills/smyx_common/scripts/dao.py (reported line 353)May include surrounding context.

python
column_names = self.__model__.__table__.columns.keys()

            for key in column_names:
                value = getattr(model, key)
                setattr(instance, key, value)

            session.commit()

The ai_chat docstring claims to invoke an agent, but the implementation never executes it.

Low
Category
Excessive Agency
Confidence
99% confidence
Finding

The docstring at L26 explicitly states that this method invokes openclaw agent through subprocess. However, the command and subprocess invocation are entirely commented out (L48–L66); the replacement empty dictionary at L68 causes attribute access to fail, and the exception handler suppresses the failure, so the method returns no analysis result.

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 · skills/smyx_common/scripts/util.py (reported line 30)May include surrounding context.

python
elif isinstance(cur, dict):
                        cur = cur.get(part, default)
                    else:
                        cur = getattr(cur, part)
                return cur
            except Exception:
                return default