Back to skill

Security audit

合同审查助手

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward contract-review prompt that reads user-provided contracts and generates a local HTML report, with no evidence of hidden execution, exfiltration, or privilege abuse.

Before installing, treat this as a legal-assistance prompt rather than a lawyer. Redact sensitive business, personal, or confidential terms where possible, confirm before processing uploaded contracts, and delete the generated contract_audit_report.html if you do not want the report retained locally.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The activation phrases are broad, generic requests such as asking to review a contract, which can easily overlap with normal conversation and unintentionally trigger the skill. Because this skill processes potentially sensitive legal documents, accidental activation could lead to users sharing confidential contract text or invoking analysis in contexts they did not intend.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The README encourages users to paste contract text or upload .docx/.pdf files but does not warn that contracts often contain highly sensitive commercial, personal, or legal information. Without a privacy notice, users may disclose confidential data without understanding retention, third-party API exposure, or downstream processing risks, especially since the tool references AI semantic analysis and an external API-backed architecture.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases are very broad, natural-language requests like '帮我审核合同' and '看看这个合同', which can cause the skill to activate in situations where the user did not explicitly intend to invoke this specific workflow. Unintended activation is especially risky here because the skill is designed to process uploaded documents and may initiate file parsing and downstream analysis on sensitive contract content.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill explicitly instructs saving the audit output as contract_audit_report.html without a clear user-facing warning or consent step for local file creation. Silent file writes can surprise users, create privacy concerns for sensitive contract data, and may leave artifacts on disk that persist beyond the session.

Content

No source excerpt is available for this finding.