Back to skill

Security audit

transit-trip-planner

Security checks for vulnerabilities and agentic risk

Overview

This transit-planning skill is purpose-aligned and disclosed, with normal external route-service data sharing but no local code, persistence, or credential access.

Before using the skill, users should understand that precise locations, travel times, route constraints, and any accessibility needs may be sent to BusMaps or Geoapify to compute routes. Use less precise endpoints when possible and review any paid-call approval prompt carefully.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T01 · Skill Instruction Hijacking

Warning
Location
SKILL.md:10
Finding

Mutable External Documentation Can Influence Agent Instructions

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
.
Finding

Sensitive Trip Details Are Submitted to External Services Without an Explicit Privacy Consent Gate

Content
View full analysis
","parameters":{...}}` to the selected listing. Preserve explicit timezone offsets; otherwise treat BusMaps times as local to the origin. Read notices and distinguish `noCoverage` from `noRouteFound`. Do not claim real-time status when live fields are absent. **Done when:** an attested route response or exact error is received. ``` ### Technical Analysis The workflow collects origins, destinations, travel times, accessibility needs, and route constraints, and then submits parameters to an external listing. Precise location and travel-time combinations can reveal a user's home, workplace, routine, planned absence, or movements. Accessibility requirements may additionally disclose sensitive mobility or health-related information. The skill includes an explicit approval requirement for potentially paid Geoapify requests, but it does not provide a corresponding privacy notice or consent requirement before transmitting itinerary data to third-party services. It also does not specify data minimization, retention expectations, or which collected fields are necessary for each operation. The reviewed file does not contain credentials or plaintext secrets. The issue concerns insufficient controls around external disclosure of user-provided trip information. ### Attack Path 1. A user provides an origin, destination, travel time, accessibility requirements, or route constraints. 2. The agent extracts these details as required by the skill. 3. The agent selects an extern ...[truncated 966 chars]
Remediation
View remediation
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep