Back to skill

Security audit

Opclawtm Skill

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly documents opclawtm setup, but it asks users to grant very broad Feishu permissions, install an unpinned global CLI, and expose local log identifiers.

Review before installing. Use a pinned and trusted opclawtm release if possible, avoid administrator/global installation unless necessary, do not paste secrets into untrusted tools, reduce Feishu permissions to the minimum needed, and avoid sharing raw gateway log output in chat.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (3)

T08 · Insecure Dependencies

Error
Location
references/installation-flow.md:32
Finding

Unpinned Third-Party CLI Is Installed Globally

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
references/feishu-config-flow.md:64
Finding

Feishu Bots Are Granted Excessive Organization-Wide Permissions

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
references/feishu-config-flow.md:288
Finding

Gateway Log Inspection Exposes Complete Sensitive Log Lines

Content
View full analysis
Remediation
View remediation
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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (14)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The entire skill reference, including direct instruction text such as line L41, is written as mandatory Chinese guidance and provides no option for the user to select another language or locale. The policy requires flagging skills that force a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The document explicitly authorizes the agent to run trial-activate and bind-license --key XXX, which are state-changing account/license operations. Allowing an agent to execute them without requiring explicit user confirmation or warning about side effects can cause unauthorized activation, license binding to the wrong account, or unintended consumption of trial/paid entitlements.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The document states the agent must not perform operations on the user's behalf, but later directs it to execute shell commands and inspect local logs. This contradiction weakens safety boundaries and can normalize broader agent access to the local environment beyond the stated purpose.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs the agent to read a local gateway log and extract ou_ Open IDs, then display them to the user. This is sensitive local-data access that goes beyond simple TUI guidance and can expose identifiers from logs without clear minimization, consent, or necessity checks.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill explicitly tells the agent to read and display Open IDs from local logs without an explicit privacy notice or confirmation. Even if Open IDs are not secrets like tokens, they are still user-linked identifiers and disclosing them from logs can violate privacy expectations and expose unrelated entries.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The document explicitly instructs the agent to execute environment and installation commands on the user's system without first requiring user confirmation or warning about system modification, package installation, and network access. In an agent context, this can lead to unauthorized changes to the host environment and normalizes auto-execution of external package manager commands, which is risky even if the commands appear routine.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The activation and license-binding flow directs the agent to contact external licensing services and handle license keys without disclosing privacy implications, what data may be transmitted, or the consequences of revealing authorization details back to the user interface. This can expose sensitive license information, device-linked activation metadata, and trigger unintended external side effects without informed consent.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

该文件中的标题和指令均直接规定以中文进行整个流程说明,但没有向用户提供语言/locale 选择,也没有说明这是仅适用于中文环境的区域性约束。根据规则,强制特定语言且无用户 opt-in 属于自然语言政策风险。

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger conditions are overly broad and include common conversational phrases like “我想要...” and “我需要一个...”, which can match ordinary user dialogue far beyond the intended requirement-analysis workflow. In this skill context, that increases the chance of unintended activation, misrouting user requests into a private-skill creation flow, and causing unauthorized or confusing task creation through the assistant workflow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The entire skill is written as prescriptive guidance in Chinese and directs the agent on exactly how to guide the user, but it does not provide any opt-in or alternative language path. Under the policy, forcing a specific language without user choice is a natural-language policy violation unless the locale constraint is clearly documented and justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

文档标题和开头指令以中文规定了整个故障排查流程的交互语境,但没有提供用户可选择其他语言的选项,也没有说明这是面向特定中文区域用户的受限工具。根据语言/地区策略,这属于未经用户选择即强制特定语言的自然语言策略问题。

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The document tells the agent to read local gateway logs for identifiers matching ou_ and then directly display the discovered user ID to the requester. That creates a natural-language exfiltration path from local logs to the chat interface, with no verification that the requester is entitled to see the identifier and no minimization of disclosed data.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

This guidance explicitly bridges sensitive local telemetry (~/.openclaw/logs/gateway.log) into user-visible output by instructing the agent to grep for ou_ IDs and return them. In the context of an agent skill, that is dangerous because any prompt that steers troubleshooting toward this branch could cause disclosure of identifiers from machine-local logs, potentially exposing another user's account data or internal workspace metadata.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

All user-directed operational instructions are presented only in Chinese, and the document does not indicate that the skill is region-specific or that users may choose another language. This can violate language or locale policy when no user opt-in or justification is provided.

Content

No source excerpt is available for this finding.