Back to skill

Security audit

Pet Pica Behavior Recognition | 宠物异食行为识别(啃咬电线/塑料)

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious because scanners report unexpected account and profile management, broad or unintended video transfers, and behavior beyond its stated pet video analysis purpose.

Consider avoiding installation until you can confirm when video is sent to the remote service, what identity and profile data is stored or transmitted, and whether export only truly prevents analysis. The scan is marked partial because one local reference was missing, so it does not establish that all referenced material 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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (25)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied chunk is a user-account database utility, not a video-analysis implementation. It can create and modify a workspace-local database containing personal information and tokens, while providing no video upload, URL processing, recognition API call, contact-duration detection, or warning logic. Although database utilities can support a larger skill, this chunk specifically implements account and credential management whose relationship to the declared purpose is not established. No network transmission or malicious use of stored information is shown; this assessment is limited to the supplied code.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The supplied code does not substantiate the declared primary purpose. The agent command is entirely commented out, so this chunk does not execute an external agent or subprocess. Imported base classes and API services could contain additional functionality, but their implementations are unavailable; this assessment is limited to the supplied chunk and does not establish what the complete skill does.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
88% confidence
Finding

The main analysis path is consistent with the description. However, --list invokes skill.get_output_analysis_list(open_id=...) and prints the returned history, adding a distinct retrieval capability beyond analyzing a supplied video. Internal identity resolution and optional result-file output are supporting implementation details, not independent mismatches. The imported analysis implementation is not supplied, so the hazardous-item detection, two-second threshold, warning behavior, and server-side processing cannot be verified from this chunk alone.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

The video submission and asynchronous result handling are consistent with the declared server-side analysis workflow. However, get_output_analysis_list() independently calls page() to retrieve existing report records, a user-facing capability not disclosed in the description. The supplied chunk does not show pet-specific detection or enforcement of the two-second threshold, but those functions could reside in the delegated API; their absence here alone is not evidence of a mismatch.

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-pet-pica-behavior-recognition-analysis"
description: "Triggers when a user provides an indoor camera video for analysis; supports local uploads or network URLs to call server-side APIs for pet pica-behavior recognition, detecting contact between the pet's mouth and non-food hazardous items (electric wires, plastic bags, socks, tissues, toy fragments, etc.); when the contact lasts ≥ 2 seconds, outputs a warning signal to help prevent intestinal obstruction, electric shock and other dangers (without diagnosing diseases). Application scenarios: indoor cameras, pet safety monitoring, smart-home security. | 当用户提供室内监控�

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 default trigger activates pet-pica analysis for any indoor surveillance video submitted for analysis.

Medium
Category
Trigger Abuse
Confidence
97% confidence
Finding

The default trigger covers any indoor surveillance video submitted for analysis, not just requests for pet-pica detection. Because the workflow sends footage to an external service, this can turn an unrelated analysis request into unintended disclosure of sensitive household video.

Content

No source excerpt is available for this finding.

The remote-analysis workflow lacks a privacy warning for indoor surveillance footage.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

Remote processing is mentioned, but the automatic workflow lacks a privacy notice or authorization checkpoint for transferring indoor surveillance footage. Such footage can include occupants, children, and private household details, making the broad automatic trigger particularly sensitive.

Content

No source excerpt is available for this finding.

The --export-env-only option promises not to analyze video, but 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 main never checks the flag. When supplied with a video input, execution still calls the analysis workflow, potentially transferring private footage despite an explicit no-analysis request.

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 an analysis API without disclosure of the upload.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The local-input branch reads the entire video into a multipart file payload and passes it to the analysis API. The provided code has no confirmation, user-facing notice, or explanatory comment that the video leaves the local system; users may not expect remote transmission of potentially sensitive recordings.

Content

No source excerpt is available for this finding.

The --export-env-only option promises not to analyze video, but execution still invokes analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The --export-env-only flag is parsed but never used to stop execution. With a video input present, the program invokes analysis despite promising not to, potentially causing an unintended external transfer of surveillance footage.

Content

No source excerpt is available for this finding.

The skill includes persistent personal-profile management unrelated to pet-video hazard detection.

Medium
Category
Excessive Agency
Confidence
88% confidence
Finding

The User model stores real names, email addresses, birthdays, sex, and age, and UserDao inherits full create, update, and delete capabilities for these records. DAO initialization also creates and migrates the user table (L180–L214). Although authentication tokens could support the declared server-side API calls, managing these personal-profile attributes is not justified by the manifest's video-analysis and warning purpose.

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 uses mode 'w', which immediately erases an existing file's contents. No docstring, explanatory comment, confirmation, or action log discloses this destructive behavior; exception tracing only runs after an error.

Content

No source excerpt is available for this finding.

Analysis requests can silently register a remote user account, exceeding the declared video-analysis workflow.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

When cached tokens are missing, an ordinary API request can invoke phoneLogin with silent=1 and register=1, causing remote account provisioning without a separate authorization step. Server-side video processing is declared, but automatic account creation and persistent identity linkage are additional side effects not disclosed in the provided workflow.

Content

No source excerpt is available for this finding.

HTTP requests can silently trigger remote account registration without disclosure.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The generic HTTP wrapper embeds account registration into an otherwise ordinary request when authentication tokens are absent. This hidden state-changing operation can transmit an identity and establish a remote account when the user has only requested pet-video analysis.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
94% confidence
Finding

This POST transmits identity values to a separate login/registration endpoint, rather than merely submitting video for analysis. More seriously, current__user_name can be assigned ApiEnum.API_SECRET_KEY and then passed here as both openId and mobile when tokens are missing, unnecessarily disclosing a credential in registration payload fields; the destination's trust and transport protections are not shown.

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 files, including during module import.

Low
Category
Output Handling
Confidence
92% confidence
Finding

When a configuration file is absent, YamlUtil.load creates its parent directories and writes the file without a warning, log, comment, or docstring disclosing this side effect. BaseEnum.init_subclass invokes this loader at L053 and L058, so defining the configuration classes during import can modify the filesystem 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:

The ai_chat docstring promises an agent invocation, but the implementation never invokes an agent or returns analysis.

Low
Category
Excessive Agency
Confidence
99% confidence
Finding

The docstring at L25–L34 states that this method invokes openclaw agent through subprocess and returns an analysis result or session ID. However, the invocation is commented out at L61–L66, and the empty dictionary at L68 causes result.stderr to raise an exception that is caught at L81–L82, leaving the method to return None without performing the documented operation.

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/util.py (reported line 30)May include surrounding context.

python
elif isinstance(cur, dict):
                        cur = cur.get(part, default)
                    else:
                        cur = getattr(cur, part)
                return cur
            except Exception:
                return default