Back to skill

Security audit

Micro-Expression Recognition & Analysis Tool | 微观情绪识别分析工具

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because scanners report behavior beyond facial-expression analysis, including identity and credential handling, remote account registration, and analysis that may still run in an export-only mode.

Treat this as a review-needed skill and consider avoiding installation until you understand which facial media and identity details go to the remote service, what account and token data is stored, and whether export-only mode can trigger analysis. The scan was partial because a referenced local path was missing, so it does not establish that all behavior was inspected.

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 (27)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied code contains only a comment and provides none of the declared functionality. This assessment is limited to the provided chunk; it does not establish what any unprovided files might implement.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The visible implementation does not deliver the declared facial micro-expression analysis capability. Its only substantive method is an incomplete generic agent wrapper; it neither executes an agent nor produces an analysis result. Imported base classes and services are not supplied, so their capabilities cannot be verified. No active subprocess execution, network transmission, or other undeclared resource access is demonstrated by this chunk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied code contains no image processing, facial-feature recognition, micro-expression analysis, or emotional-report generation. Instead, it implements user-account persistence and database schema maintenance. Database access alone could be a supporting implementation detail, but the concrete personal-profile and authentication-token management capabilities are not explained by the declared purpose. This finding is limited to the supplied module; it does not establish that the wider skill lacks emotion-analysis functionality. No network transmission or token theft is shown.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

Ordinary formatting, filesystem, and HTTP helpers could support the declared analysis function and are not independently grounds for a mismatch. However, automatic account enrollment, persistent identity management, and payment-related behavior are substantial capabilities not disclosed by the facial-analysis description. The request wrapper also attaches configured credentials without restricting absolute URLs to a trusted service origin. The supplied file contains no image processing, expression recognition, or emotional-report generation. It may support analysis code elsewhere, but that does not establish a match for the additional behaviors visible here.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
86% confidence
Finding

The primary analysis workflow matches the description, and API delegation and report-file output are supporting implementation details. However, the code also resolves an internal OpenId using a help-hidden identity argument and provides a standalone mode that retrieves analysis records associated with that identity. These account-linked capabilities are not disclosed in the declared purpose. The imported implementations are unavailable, so this assessment does not assume how identity is obtained, what media is transmitted, or whether history access is properly authorized.

Content

No source excerpt is available for this finding.

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: "emotion-analysis"
description: "Professional discernment of subtle cues! It performs detailed analysis and recognition of facial micro-expressions, outputs precise emotional state reports, and unveils true inner emotional activities. | 微观情绪(微表情)识别分析工具,专业察言观色!针对人物面部微表情进行细致分析识别,输出精准的情绪状态分析报告,揭示真实内心情绪活动"
version: "1.0.23"
license: "MIT-0"
---

# 😊 Micro-Expression Recognition & Analysis Tool | 微观情绪识别分析工具
> **智能分析中枢** · 图片/视频智能分析 · 结

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 analysis workflow lacks a privacy warning before sending facial media or its URL to a remote analysis service.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

The workflow automatically processes facial media through a cloud API without requiring a clear pre-upload privacy notice or authorization. Although remote API use is mentioned, the lack of a meaningful disclosure about sensitive facial content, account linkage, and retention can result in unintended disclosure of users' or third parties' media.

Content

No source excerpt is available for this finding.

The documented export-only option does not prevent analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The parser accepts --export-env-only and promises not to perform analysis, but the flag is never checked. With media arguments supplied, execution still initializes identity and invokes analysis, potentially transferring sensitive facial media or incurring charges despite the caller explicitly selecting a non-analysis mode.

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 submitted to the analysis API without disclosure of the upload.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The local-file branch reads the entire media file and supplies it as a multipart upload to self.analysis without a disclosure or authorization gate in this path. Facial video can expose sensitive personal information, so a caller expecting local processing could unintentionally disclose it to the analysis service.

Content

No source excerpt is available for this finding.

The documented export-only mode does not prevent video analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The export-only flag is accepted but never consulted, so supplying it alongside media input still reaches video 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.

save() is documented as creating a new record but can overwrite an existing record after an insertion failure.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

save() catches every insertion exception and then attempts to overwrite a record by ID or username/realname, copying fields including identifiers and credentials. If an untrusted or conflicting model reaches this method, an intended insertion can unexpectedly modify another existing profile; the supplied source does not establish a complete account-takeover path.

Content

No source excerpt is available for this finding.

The skill includes persistent user-profile and credential management unrelated to facial-expression analysis.

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

The User model stores identifying information and authentication material, including real names, email addresses, birthdays, token, and open_token (L448-L465). UserDao exposes persistent CRUD operations for these records (L470-L473), and DAO initialization creates and migrates the user database (L180-L215); this account-management capability is not justified by the manifest's stated micro-expression analysis and reporting purpose.

Content

No source excerpt is available for this finding.

The generic file-opening helper silently truncates existing files.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

At L101, FileUtil.open opens the supplied path in write mode, immediately erasing existing contents even before anything is written. The helper has no docstring, comment, confirmation, or operation log disclosing this destructive behavior; exception tracing only occurs if opening fails.

Content

No source excerpt is available for this finding.

The request helper silently registers remote accounts and persists returned user credentials, beyond facial-expression analysis.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

The generic request path can silently register a remote account using the current identity and then persist returned profile information and authentication tokens. This creates undisclosed identity-transmission and credential-retention side effects beyond what a user would necessarily expect from facial-expression analysis, although there is no evidence the service itself is malicious.

Content

No source excerpt is available for this finding.

An ordinary HTTP request can silently register a remote user without disclosing that side effect.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

An ordinary request can trigger a separate phoneLogin call with silent=1 and register=1, creating a remote account when credentials are missing. Hiding this state-changing identity operation inside a generic request helper prevents callers from reliably distinguishing analysis requests from account enrollment.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
87% confidence
Finding

This POST transmits the selected identity as both openId and mobile, along with its source, to a configured remote login/registration endpoint. External transmission is not inherently unsafe, but here it accompanies automatic enrollment without an explicit consent boundary in the supplied code; the actual destination and transport security cannot be verified from this source.

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"):

An insufficient-balance response redirects users into installing a payment skill and purchasing credits.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

On HTTP 402, the helper returns instructions to install smyx-payment and initiate account recharge rather than an emotion-analysis result or a neutral failure notice. This introduces an undisclosed payment and cross-skill installation workflow into a skill described solely as facial micro-expression analysis; the code instructs installation but does not itself perform it.

Content

No source excerpt is available for this finding.

Configuration loading silently creates directories and YAML files, including during module import.

Low
Category
Output Handling
Confidence
94% confidence
Finding

YamlUtil.load creates a parent directory and writes a configuration file when the requested path does not exist, without a prompt, log, docstring, or explanatory comment disclosing this side effect. BaseEnum.init_subclass invokes this loader during class creation (L053 and L058), so merely importing the module can modify the filesystem rather than only read configuration.

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 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()

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: