Back to skill

Security audit

News Briefing

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently delegates news research to CellCog, with ordinary privacy and dependency-installation considerations.

Send only information you are authorized to share with CellCog and protect your API key. Review the separate CellCog skill and SDK before installation, preferably using reviewed, pinned releases; their executable code was not included here. Confirm the notification destination and service costs, especially before premium-tier reruns.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

The usage workflow sends task prompts to CellCog without a privacy warning.

Medium
Category
Output Handling
Confidence
93% confidence
Finding

The instructions pass the user's task prompt to CellCog through client.create_chat, transmitting its contents to a third-party service. Although the skill identifies CellCog as its provider, it does not explicitly disclose that prompt contents leave the user's environment or warn against including confidential information.

Content

No source excerpt is available for this finding.

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

Medium
Category
MCP Rug Pull
Confidence
87% confidence
Finding

The installation command uses npx skills without a fixed version, allowing a mutable npm release to execute with the installing user's permissions. Although this is a skill installer rather than an MCP server, the underlying supply-chain concern is valid: a compromised future release could access local files or credentials. This is an optional setup path, and there is no evidence that the referenced package is currently malicious.

Content

No source excerpt is available for this finding.