ClawHub
Back to skill

Security audit

合同审查助手

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent contract-review helper that reads user-provided contracts and creates a local HTML report, with no evidence of hidden collection, destructive behavior, or unrelated authority.

Before installing, treat uploaded contracts as sensitive: redact personal data, signatures, seals, pricing, or trade secrets where possible, and check where the generated contract_audit_report.html will be stored or synced. The output is a review aid only, not a substitute for a lawyer on important agreements.

SkillSpector

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
Confidence
94% confidence
Finding
The trigger phrases are generic everyday requests such as asking to review a contract, which can cause the skill to activate in situations the user did not specifically intend. In a contract-review skill, accidental invocation can lead users to paste or upload sensitive legal, commercial, or personal information into the tool without clear consent boundaries.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The README explicitly encourages users to paste contract text or upload .docx/.pdf files but does not warn that these documents may contain confidential business terms, personal data, seals, signatures, or other sensitive content. This omission increases the likelihood of unintended disclosure, particularly because contract review inherently involves high-sensitivity documents.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The trigger phrases are broad, generic requests a normal user might say in many contexts, which increases the chance the skill is invoked unintentionally. Unintended activation can cause the assistant to process sensitive contract content or begin file-handling/report-generation behavior when the user did not clearly consent to using this specific skill.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill instructs the agent to save an HTML report to a local file without clearly telling the user beforehand that a file will be created or modified. Silent file creation can violate user expectations, create privacy issues if sensitive contract data is written to disk, and increase risk in environments where local files are synced, shared, or later exposed.

Static analysis

No suspicious patterns detected.