Back to skill

Security audit

Error Monitor Fix

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly aligned with OpenClaw error monitoring, but it persistently copies log-derived content into workspace files without redaction and has incomplete scope and install disclosure.

Install only if you are comfortable with OpenClaw error logs being copied into persistent workspace files. Before use, add redaction and Markdown escaping, document the dotfiles and fixed diagnostic commands, and resolve the missing postinstall script or remove that install hook.

Vulnerability Patterns
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • 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
  • 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 (2)

T02 · Agent Memory Poisoning

Warning
Location
scripts/monitor-error.js:86
Finding

Untrusted log content is written into persistent workspace memory without sanitization

Content
View full analysis
Remediation
View remediation
-]/g, '\\$&') .slice(0, maxLength); } ``` Apply this helper to every value interpolated into `lines_out`. ]]>

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/monitor-error.js:86
Finding

Potentially sensitive error-log contents are persisted and printed without redaction

Content
View full analysis
Remediation
View remediation
redact(item)); if (value && typeof value === 'object') { return Object.fromEntries( Object.entries(value).map(([k, v]) => [k, redact(v, k)]) ); } return value; } ``` ]]>
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (14)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill is presented as monitoring and recommending fixes, but the findings indicate undeclared writes to workspace files and deduplication state, while the documented output and actual behavior do not consistently match. Hidden persistence and logging behavior can leak sensitive operational data, create unauthorized local state, and mislead users about what the skill does.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill is presented as monitoring and recommending fixes, but the findings indicate undeclared writes to workspace files and deduplication state, while the documented output and actual behavior do not consistently match. Hidden persistence and logging behavior can leak sensitive operational data, create unauthorized local state, and mislead users about what the skill does.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill declares no explicit tool scope or permissions even though the documented behavior implies access to environment-dependent capabilities and local system context. In an agent framework, missing scope declarations can cause the skill to run with broader privileges than users expect, reducing transparency and increasing the chance of unintended access.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest description and main body are written as Chinese-only operational instructions, including the skill title and behavior description. Under the policy, language constraints should either offer user opt-in/choice or be clearly justified as region-specific, which is not present here.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The documentation states that repair strategies no longer perform filesystem modifications, yet it also says the skill appends to an error file. This inconsistency undermines trust and can cause reviewers or users to underestimate the write capabilities of the skill, leading to unsafe approval decisions.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

One section claims manual-only repair, while another says the skill attempts automatic fixes. In a security-sensitive automation environment, contradictory descriptions of autonomy are dangerous because they obscure whether the skill may take actions without user confirmation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The natural-language header, usage notes, and user-facing messages are written entirely in Chinese, which effectively constrains the skill's interaction language. There is no opt-in, language selection, or documented reason that this skill must be Chinese-only, so it conflicts with the locale-policy requirement.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The header claims the skill only provides manual or read-only guidance, but the implementation persists execution history to a workspace file. This mismatch can mislead users and reviewers about side effects, and the stored history may reveal operational error metadata over time.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill description suggests passive log scanning and repair suggestions, but the code executes local shell commands such as ss, grep, awk, and openclaw sessions cleanup --dry-run. Even if intended as read-only checks, invoking system commands expands the attack surface and can produce unintended effects or trust-boundary violations in an agent context.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The script stores a persistent fix history containing timestamps, error counts, and applied actions derived from recent logs without any warning or consent. Operational logs can contain sensitive system state or incident details, so persisting derived data increases exposure and retention risk if the workspace is later accessed by another process or user.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest description says the skill provides 'OpenClaw 系统级错误修复建议', which implies producing repair guidance. In the implementation, scanErrors only parses log entries, deduplicates them, prints summaries, and appends markdown rows to error.md; no logic exists to infer fixes, suggest actions, or map errors to remediation steps.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The script formats timestamps using the fixed locale 'zh-CN' and timezone 'Asia/Shanghai', which forces a specific language/locale behavior regardless of user preference or system settings. This is a natural-language policy issue because the file does not offer any locale choice or document an opt-in for this restriction.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
73% confidence
Finding

L30 describes an auto-fix.js module for automatic repair with five strategies, suggesting broader automated remediation capability. L43 then says cache cleanup and permission repair belong to other agents and are not part of this skill. This creates contradictory documentation about the breadth of remediation handled by this skill.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The description is written entirely in Chinese, which indicates a fixed language presentation without any opt-in or alternate locale. The file provides no indication that the skill is region-specific or that users can choose their language, which can violate language/locale policy requirements.

Content

No source excerpt is available for this finding.