Back to skill

Security audit

Pilot Health

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed network health helper that runs Pilot Protocol diagnostic commands, with some expected exposure of connection and known-peer metadata.

Install this if you are comfortable letting the agent run pilotctl diagnostics and view Pilot Protocol connection and known-peer metadata. Use it in environments where that network metadata is not sensitive or where the operator is authorized to inspect it.

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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (2)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The stated purpose is network health monitoring with latency and reachability checks (L003-L014), but the documented connections command reveals all active connections, ports, encryption status, and byte counts. That is broader operational introspection than the narrow health-check scope described in the manifest.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The manifest explicitly says this skill should not be used to discover new agents (L012), framing the skill as limited to health monitoring. However, the documented pilotctl --json peers command returns known agents with contact metadata, which is agent enumeration behavior and extends beyond pure latency/reachability checks.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.