Back to skill

Security audit

Direcciones AR

Security checks for vulnerabilities and agentic risk

Overview

The skill’s remote address processing, API key use, and prepaid charges are disclosed and fit its advertised purpose.

Use this service only for addresses you are authorized to share with its operator. Remove unnecessary customer identifiers and apartment details, and check its privacy and retention terms before submitting sensitive batches. Protect the API key and review checkout charges. The hosted backend’s implementation is outside the inspected 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
  • 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)

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 46)May include surrounding context.

Sin registro:

bash
curl -X POST https://direcciones-ar-api-103907796853.us-east1.run.app/v1/keys -H 'content-type: application/json' -d '{"name":"mi-agente"}'

La respuesta trae key (empieza con ak_). Guardala y enviala siempre como header Authorization: Bearer ak_.... Sin créditos la API responde HTTP 402 con checkout_url: mostrásela al usuario y no reintentes en loop.

The skill instructs the agent to send customer addresses to a remote service without a privacy warning.

Medium
Category
Output Handling
Confidence
94% confidence
Finding

The skill directs the agent to submit addresses, including potentially sensitive customer addresses in batches, to an external operator without a privacy or authorization check. Although remote geocoding is the advertised functionality and there is no evidence of malicious exfiltration, describing the underlying data as official does not establish that this intermediary is authorized to receive personal information or has appropriate retention controls.

Content

No source excerpt is available for this finding.