Security audit

protocal-agent

Security checks for vulnerabilities and agentic risk

Overview

This skill is not malicious, but it needs review because it can generate detailed biomedical procedures while automatically reading local protocol documents and sending sensitive task details to web search.

Review this before installing if your workspace may contain confidential lab SOPs, unpublished research, clinical context, or regulated biomedical work. Use it only with clear user control over which files may be read and whether web search is allowed, and do not rely on it as a substitute for institutional approvals, qualified biosafety review, clinical governance, or validated protocols.

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

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
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: 93% confidence
Finding: The skill is broadly invocable for medical and molecular biology protocols, including clinical and biomedical procedures, without meaningful scope restrictions, user qualification checks, or safety gating. In this domain, overly general triggering can enable generation of actionable wet-lab or healthcare procedure guidance for risky, regulated, or dual-use activities, increasing the chance of unsafe or inappropriate use.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill instructs the agent to read local project .md and .docx protocol files to incorporate workspace knowledge, but it does not disclose this access to the user or obtain consent. This creates a privacy and data-minimization risk because sensitive internal SOPs, unpublished methods, clinical details, or proprietary research information could be accessed and reflected in outputs unexpectedly.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill directs web searches for supplementary protocol information without warning that task details may be transmitted to third-party services. User prompts in this context may contain confidential research topics, unpublished experiments, patient-adjacent context, or sensitive biotech objectives, so silent externalization can leak sensitive information beyond the local environment.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: Path B automatically performs external searches based on the user's task name and targeted domains, again without any privacy notice or consent mechanism. Because the skill covers medical and molecular biology workflows, even the task name alone can reveal confidential R&D directions, institutional activities, or sensitive healthcare-related intents to external websites.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.