Security audit

Legal Doc Reviewer

Security checks for vulnerabilities and agentic risk

Overview

This legal document helper is mostly coherent, but it asks for broad command execution and writes or logs sensitive legal work without clear user-control boundaries.

Install only if you are comfortable letting an agent process sensitive legal documents and create local outputs. Use it in a constrained workspace, review any command execution or callback/API use before allowing it, and avoid storing privileged or confidential material in logs unless retention and access controls are clear.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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 (3)

Context-Inappropriate Capability

Medium

Confidence: 87% confidence
Finding: The skill explicitly classifies itself as requiring EXEC capability even though its documented tasks are primarily document review, diffing, summarization, and citation checking. Unnecessary command execution expands the attack surface by enabling shell invocation in a context handling sensitive legal documents, making prompt-injection or unsafe future extensions more dangerous.

Vague Triggers

Medium

Confidence: 80% confidence
Finding: The invocation boundary is vague: the skill activates based on broadly recognizing whether a request fits one of several legal workflows, without crisp inclusion and exclusion criteria. Ambiguous scoping can cause the agent to invoke the skill on unintended prompts or mixed-context inputs, increasing the chance of unauthorized file writes, over-collection of sensitive data, or unsafe use of high-privilege tooling.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill instructs writing results to local filesystem paths but does not clearly warn that it may create or overwrite files. In an agent environment, silent or implicit writes can lead to data loss, tampering with existing artifacts, or creation of misleading legal work product in predictable directories.

Static analysis

No suspicious patterns detected.