ClawHub
Back to skill

Security audit

Science & Health Research Brief (Powerdrill via Finch)

Security checks for vulnerabilities and agentic risk

Overview

This skill provides a clearly disclosed external research service, with sensible payment limits and a privacy consideration for sensitive questions.

Questions are sent through Finch to Powerdrill. Remove identifying health details and confidential research before submitting, and review the providers' privacy terms. Confirm current free-run availability; paid runs require your approval. Treat results as research information rather than medical advice.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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)

The workflow submits the user's research question to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The connected-MCP workflow directs the agent to submit the user's question to an external service without first warning about disclosure or checking for sensitive information. In this health-research context, questions may contain identifiable medical details or confidential research, creating a plausible privacy exposure even during free runs. The skill openly names the external service, which reduces concern about covert exfiltration, but its payment-approval requirement does not establish informed consent to disclose sensitive data.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.