Back to skill

Security audit

Agent Dlp

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent DLP concept, but it claims broad inspection and audit logging of sensitive agent activity without enough scoping, retention, or user-notice detail.

Review this before installing in any environment with real user data. Confirm exactly what prompts, outputs, memory entries, tool parameters, IP addresses, and MAC addresses would be logged, who can read those logs, how long they are retained, and whether sensitive values are redacted before storage.

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
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (8)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
90% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 179)May include surrounding context.

blocked, text, details = dlp.check_output("我的手机13812345678")

检查工具

result = dlp.check_tool("exec", {"command": "rm -rf /"})

text

---

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
85% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 179)May include surrounding context.

blocked, text, details = dlp.check_output("我的手机13812345678")

检查工具

result = dlp.check_tool("exec", {"command": "rm -rf /"})

text

---

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The document advertises 'transparent' and 'unnoticed' protection while also describing inspection, blocking, redaction, and logging behaviors. In a DLP skill, this can mislead operators or end users about how their content is handled, reducing informed consent and increasing the chance of unsafe deployment in privacy-sensitive environments.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest-style description says the skill is triggered by checking status, input, or output, but it does not specify concrete invocation phrases, scope boundaries, or exclusion conditions. Terms like '检查输入' and '检查输出' are broad everyday actions and could cause unintended activation without clearer constraints or negative examples.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The documentation states that the system records all operations via audit logs, but it does not warn users that sensitive prompts, outputs, or tool actions may be retained. In a DLP context, logging everything can itself become a data-exposure surface if secrets or personal data are written to logs without minimization or notice.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

A file-wide natural-language policy issue exists because the skill documentation forces a specific language presentation with no user opt-in or stated regional limitation. Under the policy, language or locale constraints should either be optional or explicitly justified.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

This markdown file states that IP addresses and MAC addresses are logged, which affects user/system privacy. The description presents this as a feature but does not include any warning or disclosure about retention, visibility, or privacy implications.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

Most user-facing instructions and descriptions are written in Chinese, but the document does not indicate that the skill is region-specific or provide an opt-in language choice. This can violate language or locale policy when a skill effectively forces a single language without user selection or justification.

Content

No source excerpt is available for this finding.