Back to skill

Security audit

Clausebreaker

Security checks for vulnerabilities and agentic risk

Overview

ClauseBreaker is a text-only legal document explainer with clear limits and privacy warnings, though users should treat its legal and translation output cautiously.

Before installing, understand that this skill may process sensitive legal documents, so redact identifiers and private details where possible. Use it to understand clauses and prepare questions, not as a substitute for a lawyer, and be especially careful with non-English documents because the skill bases its analysis on translation.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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 manifest-style description says the skill analyzes "any legal document," which is a very broad trigger scope without explicit invocation constraints or negative examples. In a manifest/markdown context, this can overlap with many ordinary user requests involving contracts or terms and may cause unintended activation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Mandating full translation of non-English legal documents before analysis can degrade meaning, especially for jurisdiction-specific legal terms where wording matters. In a legal-analysis skill, this can mislead users about rights, obligations, or enforceability because the downstream clause review is based on transformed text rather than the original language.

Content

No source excerpt is available for this finding.

Behavior Manipulation

Medium
Category
Prompt Injection
Confidence
75% confidence
Finding

Subtle instructions detected that may alter agent decision-making or introduce hidden biases.

Content

Scanner excerpt · SKILL.md (reported line 127)May include surrounding context.

md
## Rules

- **Never tell the user whether to execute or not execute the agreement.** Your job is to inform, not decide.
- **Always include the disclaimer** at the end of every analysis.
- **Be specific about jurisdiction.** Vague statements like "this may not be enforceable" are less useful than citing the specific legal principle or statute when known.
- **Keep plain-English summaries genuinely plain.** If a non-lawyer would need to Google a word you used, replace it.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The rule 'Translate first, always' hard-codes an English-first workflow for all non-English input, which can systematically introduce interpretation errors into legal review. Because the skill also discusses enforceability and statutes, mistranslation in this context is more dangerous than in ordinary summarization and may produce incorrect negotiation guidance.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.