Back to skill

Security audit

Pet Vaccination Reminder (Facial Recognition) | 宠物疫苗接种到期提醒(面部识别)

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill suspicious because scanners report behavior beyond the described pet vaccination reminder purpose, including account and data handling concerns, and the scan is partial.

Consider avoiding installation until the undeclared account, token, database, report-history, and external media handling behaviors are clarified, and confirm that sending media to the external analysis service is acceptable. The overall verdict is suspicious, not malicious. Skillspector reports partial analysis because a referenced local path was missing, so the evidence is incomplete.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/smyx_vaccination_reminder_analysis.py:60
Finding

Identity initialization discards the trusted integration sender

Content
View full analysis

Impact

Under the stated conditions, reports created for one sender are associated with the default account, and another sender's normal history request retrieves that same account's reports and report links. This is a client-side identity-isolation defect, not evidence of a remote authentication bypass.

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
98% confidence
Finding

The supplied chunk implements a general database and user-management component. Its concrete model is sys_user, with fields such as email, birthday, token, and open_token; it also adds missing user-table columns during initialization. It contains no image/video processing, pet-identity matching, vaccination schema or queries, interval comparison, or reminder output. Database access alone could support the declared workflow, but the concrete user-management functionality and resources shown are materially different. This assessment is limited to the supplied chunk and does not establish what other skill components may implement.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied chunk does not implement the declared pet-vaccination analysis workflow. Its only substantive method is a nonfunctional generic agent wrapper, and no external command actually executes. Imported base classes and API services are not supplied, so their behavior cannot be assessed; there is no evidence here of malicious execution or unauthorized resource access. The mismatch concerns the materially different visible functionality and absence of the declared core workflow.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

Generic date, JSON, file, and HTTP helpers could legitimately support the declared workflow and are not independently evidence of mismatch. However, this module additionally implements account provisioning, credential handling across unrestricted request destinations, and a payment-related response flow that the description does not disclose. These behaviors go beyond querying hospital vaccination records and returning database comparisons. The supplied chunk contains no implementation of the advertised vaccination-analysis workflow; that absence alone does not establish what the rest of the skill does. Most listed operations occur when the relevant helper methods are invoked, rather than automatically on import.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The supplied code implements a general analysis/reporting client rather than the specifically described vaccination-reminder workflow. Its report-listing method is an additional user-facing capability not disclosed in the description. No visible code identifies pets, queries hospital vaccination records, or applies the stated reminder interval. Those operations could exist in the imported API implementation, which was not supplied, so their absence here does not establish that the backend lacks them; however, the visible report-history capability remains outside the declared scope.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The strongest confirmed mismatch is the user-linked analysis-history retrieval feature, which is outside the declared vaccination-reminder workflow. The supplied code also contains no visible vaccination-specific processing: analysis is delegated entirely to skill.get_output_analysis. That imported implementation could contain the declared functionality, so its absence from this wrapper alone does not establish that the underlying service lacks it. OpenID resolution and result-file output are supporting implementation details, not independent grounds for the mismatch.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

HTTP requests, pagination, and download URLs could legitimately support the declared workflow and are not mismatches by themselves. However, this chunk exposes broader mutation interfaces and username-based user lookup without any visible connection to pet vaccination analysis. Its implemented role is a generic API service rather than the described reminder workflow. This assessment is limited to the supplied chunk: these methods are only defined here, so the code does not establish that mutations occur or that the overall skill lacks the declared functionality elsewhere.

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 analysis workflow lacks a user-facing warning that local pet images or videos are processed by an external API.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

Although the technical documentation mentions an API, the user-facing workflow does not clearly disclose that local media will leave the host. Automatic processing can therefore transmit hospital or owner-associated imagery to an external processor without an informed authorization boundary.

Content

No source excerpt is available for this finding.

The --export-env-only option promises not to run analysis, but the code ignores it and executes analysis normally.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The parser accepts --export-env-only and promises not to analyze, but the program never checks the flag. With an input supplied, execution still initializes identity and invokes analysis, potentially uploading media or incurring charges despite the caller's explicit non-analysis choice.

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
90% confidence
Finding

The method reads the entire local media file and passes it to the analysis API without a visible disclosure or consent check. In a hospital workflow, this can expose owner-associated imagery or incidental sensitive content when users expect local processing.

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 is parsed but never used to stop execution. Supplying media with this supposedly non-analysis mode still triggers identity initialization and API analysis, potentially causing unintended data transfer and service usage.

Content

No source excerpt is available for this finding.

The service implements generic remote mutation capabilities beyond vaccination-record lookup and comparison.

Medium
Category
Excessive Agency
Confidence
86% confidence
Finding

The http_put and http_delete methods forward caller-supplied URLs and arguments to HTTP PUT and DELETE operations. These mutation-oriented capabilities are not justified by the manifest's identity verification, vaccination-record retrieval, and reminder reporting purpose; this file does not establish that they are actually invoked.

Content

No source excerpt is available for this finding.

DAO initialization modifies a shared user database schema despite the manifest describing vaccination-record lookup and comparison only.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

Every DAO initialization calls both table creation and schema migration. The migration explicitly alters sys_user to add source_id and realname, introducing persistent changes to user-account storage rather than merely querying vaccination records as described in the manifest.

Content

No source excerpt is available for this finding.

The concrete UserDao exposes account modification and deletion capabilities unrelated to vaccination reminder analysis.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

UserDao inherits unrestricted record creation, modification, and deletion operations, including update_by_username at L384–L409 and delete at L411–L427. Its User model includes token and open_token fields at L460–L461, so this capability extends to modifying account credentials and deleting users, neither of which is justified by identifying pets and comparing vaccination dates.

Content

No source excerpt is available for this finding.

FileUtil.open silently truncates existing files without disclosing its write-only behavior.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The generically named open helper always uses mode 'w', which immediately erases an existing file's contents. There is no confirmation, user-facing log, docstring, or comment explaining this destructive behavior; exception reporting does not disclose successful truncation.

Content

No source excerpt is available for this finding.

The request helper silently registers user accounts and persists authentication data beyond the declared vaccination-record comparison workflow.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

An ordinary request can silently invoke phoneLogin with registration enabled, then persist returned identity information and tokens in the local database. This introduces remote account creation and durable credential storage beyond the disclosed vaccination comparison workflow, without an explicit enrollment decision.

Content

No source excerpt is available for this finding.

HTTP requests can silently trigger remote user registration without disclosure.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The authentication fallback explicitly sets silent and register to 1, so a routine HTTP call can create a remote account without a separate confirmation step. This is the same underlying enrollment issue as SDI-1 and can transmit an identity or create unwanted service relationships during a supposedly analytical workflow.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

The outbound registration request sends the selected identity as both openId and mobile; the selection can use ApiEnum.API_SECRET_KEY ahead of ordinary usernames. Consequently, a configured API secret can be copied into remote identity fields and potentially retained in account records or logs, even though the destination is a configured service rather than a proven attacker endpoint.

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 YAML files, including during module import.

Low
Category
Output Handling
Confidence
93% confidence
Finding

YamlUtil.load creates parent directories and writes a configuration file when the requested path is missing, without a log, prompt, comment, or docstring disclosing this side effect. BaseEnum.init_subclass invokes this loader during class definition (L053 and L058), so importing the module can modify the filesystem even when the caller only expects to 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 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()