Back to skill

Security audit

lgpd-data-mapping

Security checks for vulnerabilities and agentic risk

Overview

The verdict is benign with medium confidence: the scanners found no AIG issues and treated one Skillspector persistence alert as likely a false positive, though coverage is incomplete because several referenced paths were missing.

The scanners classify this skill as benign, but confidence is medium and the review is incomplete: several referenced paths were not found in the scanned artifact. Consider that limitation before installing, especially if those references matter to your use.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • 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 (1)

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

md
2. **Superfície de API e jobs**: rotas REST/GraphQL/tRPC, Server Actions, webhooks, workers e rotinas cron.
3. **Cookies, pixels e trackers web**: leitura/escrita de cookies (`httpOnly`, sessão, atribuição, marketing), `localStorage`/`sessionStorage`, Google Tag Manager, Meta/TikTok/LinkedIn pixels, scripts de replay de sessão (Hotjar, PostHog, FullStory).
4. **SDKs, eventos de produto e telemetria**: chamadas `identify`/`track` (Segment, Mixpanel, Amplitude, Firebase Analytics), APMs e error trackers (Sentry, Datadog — verificar se capturam headers, body, IP ou e-mail em breadcrumbs/logs).
5. **Coletas mobile (React Native / iOS / Android)**: permissões em `app.json`, `Info.plist` e `AndroidManifest.xml`, push tokens (Expo/FCM/APNs), identificadores de dispositivo/publicidade (IDFA/GAID), geolocalização, câmera/galeria, contatos, área de transferência (*clipboard*) e armazenamento local (`AsyncStorage` vs. `SecureStore`/Keychain).

## O que registrar por atividade de tratamento