Back to skill

Security audit

Autism Stereotyped Behavior Detection (Spinning / Hand-Flapping) | 自闭症儿童刻板行为识别(转圈/摆手)

Security checks for vulnerabilities and agentic risk

Overview

This skill needs review because its no-analysis option is ineffective and its handling of account credentials and shared identities creates privacy risks.

Wait for fixes to the no-analysis option and credential handling before using children's footage. Confirm the provider's retention policy, obtain guardian consent, and use a separate workspace for each user. Cloud processing and account creation occur during ordinary use; the current no-analysis flag does not prevent them.

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 (22)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The visible implementation does not substantiate the declared behavioral-video analysis purpose. No external agent command actually executes because the invocation is commented out. Imported base classes and services could provide additional functionality, but their implementations are not supplied, so the assessment is limited to this code chunk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The reported attachment of credentials and identity metadata to caller-controlled absolute URLs could expose API keys or authentication tokens to an attacker-controlled endpoint. Silent default-user reuse also warrants scrutiny for cross-user report access, especially because these reports concern minors' sensitive behavioral information. The document exposes an API-address override and confirms automatic identity handling, but the credential-forwarding behavior is reported by static analysis rather than independently verified from supplied source.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The reported persistence of personal information and authentication tokens creates a sensitive identity store beyond the movement-analysis purpose; linking it to children's behavioral reports increases the consequences of disclosure or account confusion. The document acknowledges automatic identity management, so identity handling itself is not concealed, but it does not explain the reported additional personal-data collection or its safeguards. The underlying implementation was not supplied, so storage protection and authorization weaknesses remain unverified.

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-autism-stereotyped-behavior-detect-analysis"
description: "Using a fixed camera in rehabilitation centers or homes, the system analyzes children's behavior videos with pose estimation and temporal action detection to recognize repetitive stereotyped behaviors, including spinning (body rotation ≥ 360°), hand flapping (non-functional repetitive arm movement), body rocking (rhythmic forward-backward or side-to-side trunk motion), etc. | 通过康复机构或家庭固定摄像头,分析儿童行为视频,利用姿态估计和时序动作检测技术识别重复性刻板动作,包括转圈(身体旋转360°以上)、摆手(手�

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 documented environment-export-only mode still executes video analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The --export-env-only flag promises not to execute analysis, but its parsed value is never checked. When supplied with --input or --url, execution still reaches skill.get_output_analysis, potentially processing sensitive videos of children despite the user's explicit non-execution choice. Any resulting upload or external disclosure depends on the delegated implementation, which is not provided.

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
88% confidence
Finding

The code reads the selected local video and passes its contents to the analysis API without a visible disclosure or consent check before transfer. Uploading video is consistent with the skill's intended function, but footage of children in homes or rehabilitation settings can reveal identities, private surroundings, and sensitive health information, making undisclosed external processing particularly consequential. The API implementation and any consent controls elsewhere are not provided, so the destination, transport security, and broader consent workflow cannot be verified.

Content

No source excerpt is available for this finding.

The --export-env-only option promises not to analyze videos, but valid video input still triggers analysis.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The parser accepts --export-env-only and explicitly promises not to perform analysis, but the program never checks this flag before calling analyze_video when valid input is supplied. This violates a user-visible no-analysis boundary and could cause unintended processing or transmission of sensitive children's behavioral videos; the actual transmission behavior depends on the unseen skill.get_output_analysis implementation.

Content

No source excerpt is available for this finding.

The documented create operation can overwrite an existing user record after an insertion failure.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

save catches any insertion exception and attempts updates by ID, then by a username-or-realname match, copying every mapped column except username in the final fallback. If an untrusted caller can influence the supplied model, a duplicate username or other insertion failure can turn an intended creation into an overwrite of another user's identifiers, tokens, and profile data without an ownership check in this path. In a child-rehabilitation setting, corruption of user identity and credentials could compromise sensitive workflows, although caller exposure and downstream token use are not shown.

Content

No source excerpt is available for this finding.

ai_chat documents an OpenClaw agent invocation, but the implementation never invokes an agent or analyzes the prompt.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The docstring at L26 explicitly claims that this method invokes openclaw agent through subprocess. However, both command construction and subprocess execution are commented out (L48–L66); the replacement empty dictionary at L68 causes attribute access at L70 to fail, and the exception handler suppresses the failure, returning None rather than the documented analysis result or session ID.

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 generically named open method always uses mode 'w', which immediately erases an existing file's contents. There is no confirmation, operation log, comment, or docstring explaining this behavior; exception tracing does not disclose successful truncation.

Content

No source excerpt is available for this finding.

The request helper silently registers remote user accounts and persists account credentials, beyond the stated video-analysis purpose.

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

An ordinary API request can automatically send the current identity to a remote login endpoint with registration enabled, then persist returned credentials through UserDao. This creates an implicit remote-provisioning and identity-disclosure side effect without an explicit authorization gate in this file; that is particularly concerning for a skill handling children's health-related information, although this code does not itself show video transmission or insecure database permissions.

Content

No source excerpt is available for this finding.

An ordinary HTTP request can silently register a remote account and persist authentication tokens without disclosure.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

This duplicates the implicit account-provisioning issue identified by SDI-2: a request helper can register an identity remotely and save authentication tokens as a hidden side effect. The concern is the missing explicit provisioning boundary, not evidence that account registration or local token storage is inherently malicious.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
87% confidence
Finding

External transmission alone is not a vulnerability, but the visible caller can select ApiEnum.API_SECRET_KEY as current__user_name and pass it to this request as both 'openId' and 'mobile'. Putting a credential into identity fields unnecessarily exposes it to account records and identity-field logging; the endpoint configuration and downstream handling are not shown, so malicious exfiltration is not established.

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 a missing configuration silently creates directories and writes a YAML file.

Low
Category
Output Handling
Confidence
93% confidence
Finding

YamlUtil.load creates parent directories and writes a configuration file when the requested path does not exist, without a log, prompt, or explanatory comment/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 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 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()

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