Back to skill

Security audit

multi-agent-orchestrator

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed multi-agent orchestrator, but it enables command execution, file writes, web access, scraping, and persistent shared state without clear user-control boundaries.

Install only if you intend to use a powerful multi-agent coordinator and are comfortable supervising agents that may write files, run commands, browse or scrape websites, and preserve shared state. Use it in a constrained workspace, review spawned-agent tasks before execution, and avoid giving it sensitive repositories or credentials unless the environment enforces its own permission controls.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill advertises and orchestrates a large set of high-risk capabilities across multiple agents, including file writes, shell/command execution, code execution, web access, scraping, and inter-agent messaging, but provides no safety boundaries, approval requirements, or user-facing warnings. In a multi-agent context, these powers compound risk because one prompt can trigger parallel or chained actions that modify the system, access external content, and persist shared state, increasing the chance of harmful or unintended operations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The agent registry uses Romanian-language role labels such as 'Analiză & Audit Date' and 'Cercetare Web' while the rest of the skill description is in English. This imposes a locale inconsistency without user opt-in or explanation, which can violate language/locale policy expectations.

Content

No source excerpt is available for this finding.