ClawHub
Back to skill

Security audit

Dual Disease Transcriptomic Ml Planner

Security checks for vulnerabilities and agentic risk

Overview

This skill is a domain-specific research-planning guide that uses public biomedical data and does not show hidden execution, persistence, or credential access.

Installers should expect this skill to provide biomedical research workflow guidance and example R package/tool recommendations. Review any generated R code and dataset choices before running them, especially because external databases, package installs, and disease-specific dataset suitability can change over time.

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
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • 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)

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · references/tissue_and_tool_decisions.md (reported line 78)May include surrounding context.

md
1. **Prefer the tissue shared by both diseases** (e.g., blood/PBMC for two systemic autoimmune diseases).
2. **If tissues are incompatible** (e.g., brain for AD + liver for NAFLD), choose the tissue of the disease with stronger GEO availability, and explicitly state that the cross-tissue comparison limits direct mechanistic interpretation.
3. **Do not mix blood-derived and tissue-derived datasets** in the same DEG analysis without batch correction and a clear statement of the limitation.
4. **Flag cross-tissue studies** in Step 7 Risk Review as a primary reviewer concern.