Back to skill

Security audit

12 Cool Skills for OpenClaw Agent

Security checks for vulnerabilities and agentic risk

Overview

The package claims to be a website vulnerability scanner, but it also bundles unrelated skills including arbitrary script execution, financial monitoring, and crypto/NFT analysis, so it should be reviewed before installation.

Install only after the publisher separates this into accurately named skills or removes the unrelated bundled capabilities. If you use the vulnerability scanner, restrict it to systems you own or are explicitly authorized to test. Do not enable the bundled script executor or indefinite monitoring features unless you have a clearly isolated environment and understand the privilege and data-exposure risks.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (2)

T05 · Unauthorized Access and Privilege Escalation

Error
Location
MASS OPENCLAW SKILL/unzip/sandbox-script-executor/SKILL.md:20
Finding

Untrusted Script Execution and Dependency Installation Without Enforced Isolation

Content
View full analysis
` | `pip3 install ` | | **Bash** | `bash
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
MASS OPENCLAW SKILL/unzip/api-test-generator/SKILL.md:20
Finding

Authorization Tokens Are Embedded Directly in Generated API Test Source

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (96)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The vulnerability-testing idea explicitly mentions SQL injection, open ports, and DevTools-related issues but only says "on your own website" informally, without enforceable authorization checks or a clear safety warning. This creates a meaningful risk that the skill could be used to probe third-party systems, facilitate offensive security activity, or normalize unauthorized testing.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

"Bot qui exécute des scripts sandbox et renvoie le résultat" describes code-execution capability without any warning, policy guardrails, input restrictions, or disclosure of what the sandbox can access. Even in a sandbox, arbitrary script execution can be abused for malware testing, prohibited automation, data handling risks, or privilege-boundary probing if the environment is misconfigured.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The documented behavior of this skill is materially different from the surrounding skill metadata/context, which claims a website vulnerability scanner while the file describes an indefinite financial/news monitoring watchdog. This kind of capability mismatch is dangerous because it can mislead reviewers and users about what the skill actually does, enabling unexpected persistent monitoring and external data collection under a false label.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill metadata provided for review says this package is a vulnerability scanner, but the embedded SKILL.md clearly defines a different capability: a code-conversion tool. This capability mismatch is dangerous because agents or users may invoke, trust, or grant permissions to the skill under false assumptions, enabling unintended behavior, supply-chain confusion, or policy bypass through mislabeled functionality.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill content is for NFT scouting, but it is packaged in the context of a vulnerability-scanner skill. This capability mismatch is dangerous because it can cause an agent to perform unintended web navigation and data collection outside the user's security-scanning intent, undermining trust boundaries and enabling skill confusion or covert repurposing. In a security toolchain, mislabeled skills are especially risky because operators may grant them elevated trust or broader access assuming they only perform defensive actions.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill content does not match the declared security-scanning context and instead performs cryptocurrency token risk analysis. This kind of capability mismatch is dangerous because it can bypass review expectations, confuse operators about what the agent will actually do, and enable deployment of an unintended skill under a trusted security label.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The file's behavior materially differs from the declared skill purpose: instead of scanning user-owned websites for vulnerabilities, it automates browsing to DexScreener and performs cryptocurrency token/rug-pull analysis. This kind of hidden capability mismatch is dangerous because it can mislead reviewers and users about what the skill actually does, enabling unauthorized or deceptive functionality under a trusted security-scanning label.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The file’s documented behavior is a generic sandbox code-execution tool, which materially conflicts with the declared parent skill context of website vulnerability scanning. This kind of capability mismatch is dangerous because it can smuggle in arbitrary code execution under a misleading trust boundary, causing operators or downstream agents to invoke a far more powerful function than intended.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill normalizes execution of user-supplied code and package installation without any warning, consent flow, or discussion of associated risks. That omission increases the likelihood that an agent will run untrusted code by default, exposing the environment to malicious scripts, dependency-based attacks, or unsafe system changes under the guise of routine task execution.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill explicitly instructs the agent to save and execute user-provided scripts and to install arbitrary dependencies, including package-manager operations. For a website vulnerability-scanning skill, this is unnecessary and introduces a powerful execution primitive that could be abused for malware staging, data exfiltration, network pivoting, or supply-chain compromise via untrusted packages.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file declares a different skill identity and purpose ('security-alerter') than the surrounding manifest context ('vulnerability-scanner'), which can cause the agent to load or invoke behavior that the user and reviewer did not expect. In an agent/tooling ecosystem, this kind of identity or scope mismatch is dangerous because it undermines trust boundaries, weakens reviewability, and can route broad security-analysis requests to unintended instructions.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

If a skill declared as a website security scanner instead performs financial/trading risk calculations, that is a strong indicator of deceptive or repackaged behavior. In an agent ecosystem, such mismatches are dangerous because they can hide unrelated data access, computations, or outbound communications behind a trusted security-testing label.

Content

No source excerpt is available for this finding.