Back to skill

Security audit

Seed Germination Rate Prediction Analysis | 种子发芽率早期预测

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because scanners report undeclared account and credential handling, remote data transfers, and behavior that does not consistently match its stated seedling-analysis purpose.

Consider holding off on installation until the publisher explains why account registration, credential storage, and user-history features are needed, and clarifies what data goes to which external service. The verdict is suspicious, not malicious; the scan was marked partial because a referenced local path was missing, so it does not establish that all behavior was reviewed.

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 manages a sys_user database with fields such as email, birthday, token, and open_token. Its operations concern user identity and local database persistence, rather than image/video processing, seedling detection, or germination-rate calculation. Persistence could support a larger application, but this chunk specifically implements unrelated user-account functionality rather than storage of seedling-analysis inputs or results. No network transmission or token exfiltration is shown. This assessment is limited to the supplied chunk.

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 substantiate the declared primary purpose. It is generic scaffolding with a disabled and broken chat wrapper, not an implemented seedling-analysis pipeline. No active subprocess invocation or other malicious behavior is demonstrated. Imported base classes and services are not supplied, so any relevant behavior they might provide cannot be verified from this chunk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

Ordinary utility functions and HTTP transport could support the declared image-analysis workflow and are not themselves mismatches. However, this chunk additionally implements an account-provisioning, credential-management, and billing workflow absent from the description or declared permissions. These behaviors occur when the relevant helpers are invoked; the snippet does not establish that they run on every use. No seedling detection, image/video processing, counting, or germination-rate calculation is present here, although a shared utility file alone cannot establish whether those functions exist elsewhere. The mismatch rests on the concrete undeclared account and credential behaviors, not merely the absence of vision code.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

The analysis entry point is broadly consistent with the declared purpose, and saving results is a supporting implementation detail. However, listing identity-linked analysis records is an additional capability not disclosed in the description. Identity resolution supports this account-scoped operation. The supplied chunk does not include the delegated analysis implementation, so it does not establish whether seedling detection and germination-rate calculation actually occur; the pet-type options alone do not prove a different primary purpose.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
84% confidence
Finding

The supplied code exposes general API and user-lookup capabilities rather than the declared analysis functionality. HTTP requests and download URLs could be legitimate supporting implementation details and are not independently evidence of a mismatch. However, the user lookup and generic mutation operations have no demonstrated relationship to the declared purpose. This assessment is limited to the supplied chunk: it does not establish that these methods are invoked, that data is exfiltrated, or that seedling-analysis functionality is absent elsewhere.

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-seed-germination-rate-prediction-analysis"
description: "Triggers when a user provides a seedling tray image or video for analysis; uses AI object detection to identify emerged seedlings (cotyledons breaking through soil or fully expanded), counts germinated seeds, and compares with total sown seeds to estimate germination rate. Application scenarios: smart seedling incubators, greenhouse nursery trays, home planting pots, seed company germination tests. | 通过育苗盘上方的固定摄像头,播种后连续采集土壤表面图像,利用AI目标检测模型识别出土幼苗,计数已发芽的种子数量,并与播种总数

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 local-media analysis workflow lacks a privacy warning about external processing and cloud report storage.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The workflow directs local-media analysis without clearly warning that the media leaves the device for external processing and associated cloud reports. References to cloud history provide some notice, but do not adequately explain media transfer or retention; greenhouse footage can contain proprietary research or incidental personal information.

Content

No source excerpt is available for this finding.

The documented configuration-only mode still executes analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The parser accepts --export-env-only and promises not to perform analysis, but the parsed flag is never checked. When supplied with media or --list, execution still reaches analysis or cloud-history operations, violating the caller's explicit no-analysis expectation and potentially causing unintended disclosure or service usage.

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
94% confidence
Finding

The local-file branch reads the complete media file and passes its bytes to the analysis service without a disclosure or authorization check at this boundary. Uploading is consistent with remote analysis, but users expecting local processing could unintentionally expose proprietary seed-test footage or incidental personal data.

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-env-only flag is accepted but never used to stop execution. With a media input or --list, a caller explicitly requesting configuration-only behavior can still trigger remote processing or history retrieval, potentially exposing data or incurring service usage.

Content

No source excerpt is available for this finding.

The service provides username-based user lookup, which is unrelated to seedling detection or germination-rate estimation.

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

At L098-L101, get_user_by_username delegates a supplied username to a user-lookup utility. Looking up user accounts is not an evident requirement of counting emerged seedlings or comparing that count with total sown seeds, and the manifest does not declare user-directory functionality.

Content

No source excerpt is available for this finding.

The skill includes user-profile and credential database management unrelated to seed-germination analysis.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

L448-L465 define a system-user table containing personal information and authentication tokens, while UserDao at L470-L473 exposes inherited read, update, and delete operations on those records. DAO initialization also creates and migrates this table (L180-L215), introducing account-management side effects beyond the stated purpose of detecting seedlings and estimating germination rates.

Content

No source excerpt is available for this finding.

ai_chat claims to invoke the OpenClaw agent, but agent execution is disabled and the method cannot return an analysis result.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The docstring explicitly states that this method invokes openclaw agent through a subprocess and returns an analysis result or session ID. However, command construction and subprocess execution are commented out; the method instead assigns an empty dictionary to result, accesses its nonexistent stderr attribute, and catches the resulting exception, returning 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

Opening a path with mode 'w' immediately erases any existing file contents, even before the caller writes anything. This generically named helper has no docstring, comment, confirmation, or user-facing log explaining that destructive behavior; its exception handler only runs after a failure.

Content

No source excerpt is available for this finding.

Analysis requests can silently provision a remote user account and persist account credentials, beyond the declared seed-germination analysis purpose.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The request wrapper can automatically register a remote account with silent=1 and register=1, then persist returned authentication tokens without a consent gate in the shown flow. This creates remote account and credential-storage side effects that users requesting seed-germination analysis would not reasonably infer from the description.

Content

No source excerpt is available for this finding.

Automatic remote account registration transmits the user's identity without disclosure.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

Automatic registration sends the selected identity as both openId and mobile without an explicit disclosure or consent check in this code. The selected identity can originate from configured user information or API_SECRET_KEY, creating an unexpected privacy disclosure and potentially transmitting a secret as account metadata.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
93% confidence
Finding

This POST is the transmission sink for implicit account registration, sending identity and source metadata to the configured health-service endpoint. External networking is not inherently unsafe, but here the payload can contain a configured API secret repurposed as an identifier, and no consent check is shown; the actual destination and its trustworthiness cannot be determined 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"):

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

Low
Category
Output Handling
Confidence
92% confidence
Finding

YamlUtil.load creates missing parent directories and writes a configuration file without a log, confirmation, or explanatory comment/docstring. Because BaseEnum.init_subclass invokes this loader when the enum classes are defined (L044–L059), importing the module can modify the filesystem even though the apparent operation is configuration loading.

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: