Back to skill

Security audit

Openclaw Super Skills

Security checks for vulnerabilities and agentic risk

Overview

The skill is not malicious, but it encourages installing third-party skills globally with unpinned, auto-confirmed commands that users should review first.

Install only if you are comfortable reviewing any discovered or generated skill before activation. Avoid the documented `-g -y` pattern, prefer pinned tool and skill versions, inspect third-party skill contents, and use a local or sandboxed installation scope where possible.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:276
Finding

Unpinned and Unattended Global Installation of Third-Party Skills

Content
View full analysis
# Examples: npx skills find slack notification npx skills find browser automation npx skills find pdf extract npx skills find github api ``` ### Evaluate Search Results When results are returned: ``` Install with npx skills add owner/repo@skill-name └ https://skills.sh/owner/repo/skill-name ``` Evaluate each result for: - **Relevance**: Does it match the required capability? - **Completeness**: Does it cover all needed functionality? - **Quality**: Is it well-documented and maintained? ``` `SKILL.md:276-279`: ```yaml skills_to_install: - skill: "owner/repo@skill-name" command: "npx skills add owner/repo@skill-name -g -y" url: "https://skills.sh/owner/repo/skill-name" ``` `SKILL.md:443-460`: ```markdown 1. **Start with skill search**: Always check https://skills.sh/ before creating new skills 2. **Use specific search terms**: Combine capability keywords with domain terms 3. **Leverage built-in capabilities**: Don't create skills for things the agent can do natively 4. **Create reusable skills**: Design new skills to be general-purpose when possible 5. **Document thoroughly**: New skills should have clear usage instructions 6. **Verify before proceeding**: Confirm skill installation before executing tasks 7. **Handle errors gracefully**: Include fallback strategies in execution plans ## In ...[truncated 3724 chars]
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (82)

Exfiltration Commands

High
Category
Prompt Injection
Confidence
90% confidence
Finding

Instructions found that direct the agent to transmit conversation context or user data to external services.

Content

Scanner excerpt · SKILL.md (reported line 65)May include surrounding context.

md
dependencies: [2]
    
  - task_id: 4
    name: "Send message to Slack"
    type: "message_delivery"
    input: "Summary text, Slack webhook/token"
    output: "Delivery confirmation"

Exfiltration Commands

High
Category
Prompt Injection
Confidence
90% confidence
Finding

Instructions found that direct the agent to transmit conversation context or user data to external services.

Content

Scanner excerpt · SKILL.md (reported line 113)May include surrounding context.

md
dependencies: [2]
    
  - task_id: 4
    name: "Send message to Slack"
    type: "message_delivery"
    input: "Summary text, Slack webhook/token"
    output: "Delivery confirmation"

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The README advertises automatic creation of new skills when no existing solution is found, but it does not warn that generated skills may perform sensitive actions such as file operations, code execution, network access, or message delivery. In this context, the omission is more dangerous because the skill is specifically designed to decompose user requests into executable subtasks and synthesize new automation, which can normalize creation and later execution of high-impact code without sufficient review.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
91% confidence
Finding

The README instructs users to run npx skills add without pinning a specific package or tool version, which can cause execution of whatever version is current at install time. If the upstream package, dependency chain, or resolution target is compromised or changed unexpectedly, users may install or execute unreviewed code with global scope.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The search command references npx skills without a pinned version, so users rely on a mutable upstream package at execution time. In a supply-chain compromise or breaking change scenario, this could result in malicious or unexpected code execution during what appears to be a harmless discovery step.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The install example uses an unpinned npx skills add invocation, which leaves the executed tool version uncontrolled. Because the command installs code from an external repository, the combination of mutable tooling and external content increases supply-chain risk and can expose users to malicious package behavior or altered install semantics.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The initialization example again depends on an unpinned npx skills execution path, which makes behavior dependent on the latest resolved package version. That creates avoidable supply-chain and reproducibility risk, especially in a skill whose purpose includes generating additional executable artifacts.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The installation instructions use -g -y, which can modify the user's global environment and bypass confirmation prompts, but the README does not warn about those side effects. In practice this reduces user scrutiny and can make accidental or malicious system-wide changes more likely.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
90% confidence
Finding

The README instructs users to run npx skills without pinning a specific package version, which allows whatever version is currently resolved from the registry to execute. That creates a supply-chain risk: a compromised latest release or dependency could run arbitrary code during install or use, especially since the command is presented as a standard setup step.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
87% confidence
Finding

The documented npx skills find <关键词> command relies on an unpinned package resolution, so users may execute an unexpected or malicious package version from the registry. In a skill ecosystem context, normalizing unpinned execution increases exposure to supply-chain compromise.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
92% confidence
Finding

The README recommends npx skills add <owner/repo@skill> -g -y, combining unpinned package execution with global installation and automatic confirmation. If the resolved package or fetched skill is malicious, this can lead to arbitrary code execution and persistent system-wide changes without an interactive review step.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
86% confidence
Finding

The npx skills init <技能名称> example again invokes an unpinned remote package, exposing users to execution of an unintended version. While initialization is lower risk than global installation, it still executes code from the registry and may create files or templates under attacker control if the package is compromised.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill description says it should be used for complex multi-step requests, automation workflows, or breaking down large tasks, which is broad enough to overlap with many ordinary user requests. Over-broad activation increases the chance the skill will trigger in contexts involving sensitive data, external integrations, or code execution pathways that were not specifically requested.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The instruction 'When receiving a user request, follow these steps' is ambiguous because it does not define which requests should invoke this skill. In an agentic environment, ambiguous activation can cause the skill to run on unrelated prompts and initiate external tool discovery or installation workflows unexpectedly.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The skill instructs the agent to run npx skills find <keyword> without pinning a package version. npx will fetch and execute the latest package code from the registry, creating a supply-chain risk where a compromised or malicious update could run arbitrary code in the agent environment.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

This unpinned npx skills invocation executes remote package code with whatever version is current at runtime. That makes the skill's behavior non-deterministic and exposes the agent to remote code execution if the package or one of its dependencies is compromised.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The example command uses npx to resolve and execute an unpinned package from the network at runtime. In an agent skill, that can become an implicit code execution path controlled by upstream publishers rather than the skill author or deployer.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Running npx skills find pdf extract without a pinned version allows arbitrary package updates to alter behavior or introduce malicious code. Because the skill encourages repeated CLI usage during decomposition, the exposure is multiplied across many agent actions.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The unversioned npx skills find github api command creates a supply-chain execution surface each time the skill searches for capabilities. If the package is replaced, typo-squatted, or its dependencies are compromised, the agent may execute attacker-controlled code.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
94% confidence
Finding

The installation guidance npx skills add <owner/repo@skill> relies on an unpinned skills package, so even the installer itself is fetched dynamically. This combines package supply-chain risk with installation of third-party skills, increasing the chance of executing untrusted code or content.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Using npx skills init <skill-name> without a fixed version means code generation depends on whatever package version is current at runtime. A malicious release could write unsafe scaffolding, embed backdoors, or modify local files unexpectedly.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The execution plan includes npx skills add owner/repo@skill-name -g -y, which both dynamically fetches the CLI and non-interactively installs a skill. This creates a high-risk chain where remote package code and remote skill content may be accepted automatically without human review.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

The sample search command uses an unpinned npx package invocation. In the context of an agent skill that may act autonomously, this can convert a documentation example into a repeated remote code execution primitive.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

Another unpinned npx skills example appears in a workflow decomposition scenario. Because the skill normalizes these commands as standard practice, it increases the likelihood that downstream users or agents will execute untrusted package code as part of ordinary task planning.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
95% confidence
Finding

This example again directs use of npx skills find academic search without pinning. Repeated occurrences indicate the unsafe pattern is systematic, not incidental, making exploitation easier if the package ecosystem is compromised.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.