Back to skill

Security audit

Data Analysis

Security checks for vulnerabilities and agentic risk

Overview

The skill openly uses CellCog to analyze selected datasets, with ordinary privacy and software installation considerations.

Upload only datasets you are authorized to share with CellCog, and remove unnecessary personal information and secrets first. Check its retention and deletion terms, and review the separate CellCog dependency before installation. The full scanner report was complete for the sole bundled Markdown file; this review also independently inspected that entire file.

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
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
Findings (2)

Dataset upload instructions lack a warning about sharing sensitive data with a third-party service.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The skill encourages uploading datasets to an external analysis service without a sensitive-data warning or an explicit authorization checkpoint. Examples involving customer records, HR salaries, server logs, and SQL dumps increase the likelihood of disclosing personal information, credentials, or confidential business data; external processing is an advertised feature, not evidence of covert exfiltration.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx skills'.

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

The command npx skills add cellcog/skills --skill cellcog can download and execute an unpinned npm package, while the skill source is also not tied to an immutable revision. A compromised package release or changed skill repository could introduce malicious behavior with the agent's local permissions; this is a supply-chain risk, although the static finding incorrectly describes the command as an MCP server reference.

Content

No source excerpt is available for this finding.