Security audit

san-sheng-liu-bu

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed, instruction-only workflow skill for structured multi-agent task planning, with expected local session notes but no hidden scripts, credentials, or network behavior.

Install this if you want a structured multi-agent planning and execution workflow. Expect it to create .court-session records in the current project and to delegate larger tasks across multiple agents, sometimes in the background. Use it in a controlled workspace for sensitive repositories, and avoid the quick-path command when you want the review step preserved.

SkillSpector

Vulnerability Patterns

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

Confidence: 82% confidence
Finding: The trigger conditions are broad enough to activate on many ordinary task requests, causing the skill to engage more often than users may expect. In this skill, over-triggering is more dangerous because activation leads to a multi-agent workflow that can create files and orchestrate execution steps, increasing the chance of unintended side effects.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The skill instructs the agent to create directories and files in the current working directory without requiring an explicit user-facing warning or confirmation. This can lead to unexpected filesystem modifications, repository pollution, accidental inclusion of sensitive task content in project files, or writes in contexts where the current directory is security-sensitive.

Static analysis

No suspicious patterns detected.