Back to skill

Security audit

Hydroponic Nutrient Concentration Visual Assessment | 水培植物营养液浓度视觉评估

Security checks for vulnerabilities and agentic risk

Overview

The scanners rate this skill suspicious because its reported account, credential, and data-transfer behavior extends beyond the stated hydroponic image assessment purpose, and some actions may happen without clear user notice.

Consider avoiding installation until the account registration, credential use and storage, and external media processing are clearly explained and acceptable to you. The scan is partial: two local references were missing, so the evidence does not cover those referenced items.

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 (26)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied chunk is a user-database utility rather than a hydroponic imaging or diagnostic implementation. It defines a sys_user table containing names, email, birthday, demographic fields, and tokens, and provides methods to query, insert, update, and delete records. Although database storage can support a larger application, this chunk's user-account and token-management behavior has no demonstrated connection to the declared plant-analysis purpose. No camera capture, image analysis, nutrient assessment, or adjustment recommendations appear here. The code does not show token exfiltration or external network activity; the mismatch concerns its unrelated local persistence capabilities.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

Ordinary formatting, filesystem helpers, and HTTP transport could support the declared skill and are not independently evidence of mismatch. However, this chunk additionally implements persistent identity management, silent remote registration, credential-bearing requests, and payment-related messaging that the description does not disclose. No camera capture, image analysis, plant-health classification, or nutrient-adjustment logic appears in the supplied chunk. This assessment is limited to the supplied code; it does not establish that those advertised functions are absent elsewhere or that the observed credential handling is malicious.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied code presents a generic video-analysis and history-retrieval interface. Its analysis operation delegates directly to skill.get_output_analysis(input_path), with no visible hydroponic-specific instructions or processing. The imported implementation might provide specialized analysis, but that cannot be established from this chunk. Video input alone could support the declared workflow; however, the generic interface and explicit user-associated history retrieval extend beyond what the description represents. There is no evidence here of malicious behavior or access to fixed cameras.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The supplied chunk implements a general backend API interface and contains no camera capture, image analysis, plant-health classification, or nutrient-adjustment logic. HTTP utilities could support the declared workflow, so networking alone is not a mismatch; however, the exposed user-lookup and generic management functionality is not explained by the description. This assessment is limited to the supplied chunk: it does not establish whether other files implement the declared functionality, and it shows no automatic requests or malicious data transfer.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
84% confidence
Finding

The supplied code implements a generic media-analysis and report-management client. It contains no visible camera integration, plant-specific analysis configuration, root-color or leaf-morphology interpretation, or nutrient-adjustment logic. Delegating AI analysis and polling for results could support the declared purpose, but this chunk does not establish that its analysis service performs that purpose; its explicit video inputs and generic report operations instead indicate a broader workflow. Imported API implementations are unavailable, so their underlying analysis capabilities cannot be verified.

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: "smyx-hydroponic-nutrient-assessment-analysis"
description: "Using fixed cameras on a hydroponic growing rack to capture high-resolution images of plant roots (in transparent containers) and leaves (young and old), AI vision analysis identifies root color (white = healthy, yellow = early stress, brown = severe stress, black = rotting) and leaf morphology (tip burn, leaf-margin scorch, yellowing, curling) to judge whether the nutrient solution is too concentrated or too dilute, and. | 通过水培种植架的固定摄像头拍摄植物根系(透明容器)和叶片(新叶、老叶)的高清图像,利用AI视觉分析技术识别根

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 does not warn users that local plant images or videos are processed by an external API and associated with a persistent report identity.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

Although the document mentions APIs, cloud history, and persistent identity, it does not clearly tell users before analysis that local media contents leave their device and become associated with that identity. Images from homes, research facilities, or commercial growing operations can contain sensitive background information, making implicit upload authorization a privacy risk.

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 proceeds with analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The export-only argument is parsed but never checked, so supplying it with an input still reaches the analysis workflow. A user expecting no analysis can therefore trigger external media processing and associated account or billing effects without intending to authorize them.

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
94% confidence
Finding

The method reads the entire selected local file and passes its contents to the analysis API without an upload notice or authorization check at this boundary. Plant footage may also expose people, facility layouts, or proprietary growing operations, so assuming that local analysis implies consent to external transfer creates a privacy risk.

Content

No source excerpt is available for this finding.

The documented export-only mode still executes analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The export-only flag promises not to execute analysis but is never consulted after parsing. When media arguments accompany it, the program initializes identity and invokes analysis, potentially transferring private footage despite the user's explicit non-analysis selection.

Content

No source excerpt is available for this finding.

The service exposes username-based user lookup, which is unrelated to hydroponic image assessment.

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

At L098–L101, get_user_by_username delegates a caller-supplied username to a user-lookup utility. Looking up user accounts is a separate capability from analyzing plant images and recommending nutrient adjustments, and the manifest does not declare account-management functionality.

Content

No source excerpt is available for this finding.

The skill includes user-profile and credential database management unrelated to hydroponic image assessment.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The User model stores personal information and token/open_token credentials, and UserDao exposes inherited operations to read, create, update, and delete these records. DAO initialization also creates the registered tables and migrates sys_user (L180-L215). Managing user identities and credentials is not justified by the manifest's purpose of assessing plant images and recommending nutrient adjustments.

Content

No source excerpt is available for this finding.

ai_chat documents an agent invocation but never invokes the agent or produces an analysis result.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The docstring at L26 explicitly says this method invokes openclaw agent through subprocess and promises an analysis result or session ID at L33–L34. However, command construction and execution are commented out; L68 instead creates an empty dictionary, whose stderr access raises an exception that is caught, leaving the method to return None without performing analysis.

Content

No source excerpt is available for this finding.

FileUtil.open silently truncates existing files without disclosing its destructive behavior.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The generic open helper always uses mode 'w', which immediately truncates an existing file even if the caller never writes anything. There is no docstring, comment, confirmation, or user-facing message explaining this destructive behavior; exception reporting does not disclose successful truncation.

Content

No source excerpt is available for this finding.

Nutrient-analysis requests can silently register backend accounts and persist user credentials, beyond the described image-assessment purpose.

Medium
Category
Excessive Agency
Confidence
91% confidence
Finding

An ordinary HTTP request can trigger a backend login request with silent registration enabled and then persist returned tokens and user information locally. These identity-management side effects are not disclosed by the hydroponic assessment description and can cause unexpected account creation, identity transmission, and credential retention.

Content

No source excerpt is available for this finding.

The HTTP helper silently performs remote login or registration and persists returned credentials.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The generic HTTP helper implicitly performs remote account provisioning and saves authentication material rather than merely sending the requested assessment call. This is the same underlying issue as SDI-1: callers can incur undisclosed identity and credential-storage side effects without a separate consent step.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

The login payload sends the selected identity as both 'openId' and 'mobile'; upstream selection prioritizes API_SECRET_KEY, so the helper can transmit an API secret as account-profile data to the health backend. Even if that backend is legitimate, repurposing a secret as identity data unnecessarily exposes it to registration processing, logging, or storage beyond its intended authentication use.

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 writes a configuration file when the path is missing.

Low
Category
Output Handling
Confidence
91% confidence
Finding

YamlUtil.load performs filesystem writes without a prompt, log, docstring, or explanatory comment disclosing this behavior. BaseEnum.init_subclass calls this loader during class creation (L053 and L058), so merely importing the module can create configuration files without an explicit save request.

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: