Back to skill

Security audit

flights-overhead

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its flight-identification purpose, but it may use a saved home address for a third-party flight lookup without clearly requiring fresh confirmation.

Review this before installing if you care about location privacy. The skill should be used only with a location you intentionally provide or confirm, preferably coarse city or neighborhood-level location when sufficient. Be cautious about allowing it to read a saved home address automatically, and note that the referenced flight script was not included in the reviewed artifact.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: "flights-overhead"
description: "See what flights are above you right now. Identify contrails, planes overhead, and nearby air traffic using OpenSky Network (free, no key)."
---

# Flights Overhead Skill

See what flights are currently above you. Identify contrails, planes overhead, and nearby air traffic using OpenSky Network (free, no API key needed).

## When to Use

- "What flights are above me right now?"
- "What plane is that?"
- "What's that contrail?"
- "What flights are near me?"
- "What's flying overhead?"
- Any query about identifying aircraft in the sky

## Requirements

- **OpenSky Network** — fre

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The phrase "Any query about identifying aircraft in the sky" is a broad activation condition rather than a specific trigger list. It does not clearly bound what should or should not invoke the skill, which increases the chance of unintended activation for general aviation discussion.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The document's privacy section says the agent should confirm or ask for the user's location, but the later procedure authorizes using home address data from USER.md as a fallback. This inconsistency can cause agents to bypass meaningful consent and use stored sensitive location data in ways the user may not expect for a real-time flight lookup.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs the agent to use a home address from USER.md as a fallback for obtaining location, which accesses broader personal data than is necessary for the immediate task. Even if intended as a convenience, this creates unnecessary collection and transmission of sensitive location data to a third-party API without an explicit just-in-time user confirmation for the current request.

Content

No source excerpt is available for this finding.