Security audit

Auto Research Pipeline

Security checks for vulnerabilities and agentic risk

Overview

The skill is transparent about automating research and multi-agent project work, but it asks for broad autonomous action without enough user control or boundaries.

Review this skill carefully before installing. It is not showing theft, deception, or persistence, but it is designed to let an agent run experiments and coordinate other agents automatically; use it only in a workspace where broad automated code changes are acceptable and where AutoResearch setup clearly limits editable files and commands.

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

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

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill explicitly chains fully automated experimentation, analysis, and multi-agent project delivery into a single invocation without any warning, approval checkpoint, or scope constraint. This is dangerous because it can trigger broad autonomous code changes and follow-on agent actions that exceed user intent, increasing the risk of unsafe modifications, destructive commands, or large unintended project changes.

Natural-Language Policy Violations

Low

Confidence: 94% confidence
Finding: The skill metadata description is written only in Chinese, which imposes a specific language choice without indicating user selection or multilingual support. Under the policy, language constraints should be opt-in or clearly justified.

Static analysis

No suspicious patterns detected.