Back to skill

Security audit

Flowering Date Prediction | 开花植物花期预测

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious with high confidence because scanners report substantial account, identity, and media-analysis behavior that does not match its stated flower-date prediction purpose.

The scanner recommendation is DO_NOT_INSTALL, and the verdict places this skill in the suspicious review bucket. Before considering installation, seek clarification from the publisher about why it manages accounts and tokens, sends identity or media to remote services, and exposes pet or health analysis functions; also ask how the ignored export-only option and undeclared tool scope are addressed. Because the scan was partial and some imported behavior could not be assessed, the evidence is incomplete.

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 chunk is a user-management database module, not a flowering-date prediction implementation. Although generic database access could support the declared skill, the concrete sys_user schema and account/token operations have no demonstrated connection to flower imagery, environmental measurements, or phenology inference. No camera/drone input, GDD or light processing, model inference, or bloom-date prediction appears in this chunk. This establishes a mismatch in the supplied code, without implying that other unseen files lack prediction functionality or that stored tokens are exfiltrated.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The visible implementation does not substantiate the declared flowering-date prediction capability. It is generic scaffolding with an incomplete chat helper. No external command is actually executed because the subprocess code is commented out. Imported base classes and API services are not supplied, so their behavior cannot be assessed; the mismatch finding is limited to the visible code.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied code identifies itself as a video-analysis tool and exposes video-input and analysis-history operations. It contains no visible handling of flower-bud stages, temperature or light measurements, GDD calculations, phenology models, or bloom-date prediction. The underlying analysis implementation is imported and not supplied, so its exact processing cannot be verified; however, the visible user-specific history capability is undeclared, and the exposed workflow materially differs from the stated agricultural prediction workflow.

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 HTTP helpers could support the declared application and are not mismatches by themselves. However, this chunk additionally implements an undeclared account-registration, identity-persistence, authentication, and billing integration that accesses local identity resources and a remote login service. No image acquisition, environmental-data processing, phenology inference, or bloom-date prediction is present in the supplied chunk. As this is a shared utility file, that absence alone does not establish that the complete skill lacks prediction functionality; the mismatch is grounded in the concrete additional identity and account behaviors shown.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding

The supplied code implements an API-backed media-analysis/reporting client. It contains no explicit flower-bud stage processing, GDD or accumulated-light handling, phenology-model invocation, or bloom-date prediction logic. It also handles healthAiResponse outputs and lists prior reports. File validation, uploading, and polling could support the declared purpose, but the visible workflow and response types are substantially more generic and partly unrelated. Because the API implementation is not supplied and arbitrary parameters can be forwarded, flowering prediction could exist downstream; it is not established by this chunk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

Generic AI-analysis endpoints could support the declared prediction workflow and are not inherently a mismatch. However, the explicitly configured '/health/order/api/getReportDetailExport?id=' resource concerns health-order reports, which is inconsistent with the declared flowering-date prediction purpose. This chunk only defines configuration; it does not demonstrate actual network requests or report exports. The absence of prediction logic alone is not decisive because this is a configuration file.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
86% confidence
Finding

HTTP wrappers, pagination, and download URLs could legitimately support the declared skill and are not inherently mismatches. However, this chunk also exposes an explicit user-directory lookup and general backend management capabilities without a visible connection to flowering prediction. Its visible role is a reusable backend API service, not the described prediction workflow. This assessment is limited to the supplied chunk; it does not establish malicious activity or show that these methods are actually invoked.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
86% confidence
Finding

The visible interface contains pet-domain functionality not disclosed by the flowering-date description: --pet-type accepts cat/dog/other, modifies DEFAULT__PET_TYPE, and --list retrieves identity-linked analysis records under a pet-health label. These indicate a mismatch beyond ordinary output handling. The actual analysis is delegated to imported code, so this chunk does not establish whether flowering prediction is implemented downstream. It also shows no explicit environmental-data inputs, GDD/light processing, or phenology inference, but those omissions alone would not prove a mismatch because the imported service could implement them.

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-flowering-date-prediction-analysis"
description: "AI-powered flowering-date prediction for ornamental/cut-flower plants. From fixed greenhouse cameras or drones, captures images of flower-bud developmental stages, combines environmental sensor data — cumulative temperature (Growing Degree Days, GDD) and accumulated light (PAR or daylight hours) — and uses a pre-trained phenology model to predict the full-bloom date within the next 3-7 days. Helps growers precisely schedule pollination, harvesting and tourism activities. Scenarios: smart-agriculture greenhouses, cut-flower production bases, botanical gardens, flower tourism p

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 clear privacy warning that local images or videos are sent to a remote API.

Medium
Category
Output Handling
Confidence
89% confidence
Finding

The skill automatically processes local attachments without clearly warning that their contents are uploaded to a third-party service. Although API use and cloud reports are mentioned, greenhouse or drone imagery can expose workers, facility layouts, location metadata, or commercially sensitive crop information without informed authorization.

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 analysis when input is provided.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The parser accepts '--export-env-only' and promises no analysis, but execution never checks the flag before initializing identity and analyzing supplied media. A caller relying on this non-analysis mode can unintentionally trigger remote processing, disclose imagery, or incur service charges.

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

The --export-env-only option promises not to run analysis, but the code ignores it and still processes the supplied input.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The export-only flag is parsed but never consulted, so supplying input still invokes the media-analysis workflow. This violates the caller's explicit non-analysis expectation and can cause unintended remote disclosure or service usage.

Content

No source excerpt is available for this finding.

The flowering-prediction skill includes unrelated user-profile and credential database management.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The User model stores personal information and token/open_token fields, and UserDao exposes the inherited database read, update, and delete operations for these records. Dao initialization also creates registered tables and explicitly migrates sys_user (L184-L215), introducing user-account database changes unrelated to the manifest's image, environmental-sensor, and flowering-date prediction functions.

Content

No source excerpt is available for this finding.

ai_chat documents invoking the OpenClaw agent, but agent execution is disabled and no analysis result is produced.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The docstring at L25–L34 explicitly promises a subprocess invocation of the OpenClaw agent and an analysis result or session ID. However, the command and subprocess invocation are commented out; the replacement empty dictionary at L68 causes result.stderr at L70 to raise an exception, which L81–L82 suppress, so the method returns None without performing the documented 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
97% confidence
Finding

The generically named open helper opens files in write mode, immediately erasing existing contents even before the caller writes anything. No docstring, comment, confirmation, or user-facing message explains this behavior; exception tracing only runs after an error.

Content

No source excerpt is available for this finding.

The prediction skill includes automatic remote account registration and persistent identity management unrelated to its declared phenology-analysis purpose.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

An ordinary HTTP operation can automatically register a remote account using the current identity and persist returned profile information and credentials. This introduces undisclosed privacy and account-management side effects beyond the flowering-prediction purpose, without an explicit consent gate in the shown implementation.

Content

No source excerpt is available for this finding.

HTTP requests can silently trigger remote account registration and identity transmission.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

The authentication fallback submits silent=1 and register=1 to a remote login endpoint without a separate user approval step. A user requesting agricultural analysis can therefore trigger identity transmission and account creation merely by making a normal service request.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
92% confidence
Finding

The POST transmits the selected identity as both openId and mobile to a configured remote service during automatic registration. The selected value can originate from an API secret or current user identity, creating an unnecessary disclosure risk; this is not evidence that the destination itself is malicious.

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.

Low
Category
Output Handling
Confidence
93% confidence
Finding

YamlUtil.load creates the parent directory and writes YAML when the requested file is missing, without a warning, log, comment, or docstring explaining this side effect. BaseEnum.init_subclass invokes this loader during class definition (L053 and L058), so simply importing the module can modify the filesystem.

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: