Security audit

subagent-testing

Security checks for vulnerabilities and agentic risk

Overview

This markdown-only skill gives a coherent testing methodology for validating skills in fresh agent sessions, with no hidden execution, persistence, or data access behavior found.

Before installing, be aware that this skill may activate for generic testing or validation requests; narrowing the triggers would reduce noise. Review any separately installed Claude Code plugin from the linked external repository on its own, because it is not covered by this artifact review.

A.I.G

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

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 (1)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The trigger list includes broad terms like 'testing', 'validation', and 'subagents' that can match many ordinary developer requests unrelated to this specific skill. That can cause unintended auto-activation, injecting extra instructions into unrelated workflows and increasing prompt-surface area, though the skill content itself is not overtly dangerous.

Static analysis

No suspicious patterns detected.