Back to skill

Security audit

ent-zalo-connect

Security checks for vulnerabilities and agentic risk

Overview

The skill is purpose-aligned for connecting a Zalo bot, but it automatically stores a bot token, changes gateway environment configuration, restarts the gateway, and sends a live message with only weak confirmation boundaries.

Install only if you are comfortable letting the agent save a Zalo bot token into gateway environment configuration, restart that gateway, and send a real test message to the specified chat. Treat the token as a secret, confirm the target chat before running setup, avoid pasting tokens in quoted examples or unrelated discussion, and rotate the token if it may have been exposed in chat or logs.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill defines broad trigger phrases like 'Kết nối Zalo Bot giúp tôi' and 'Cấu hình Zalo Bot: token ..., nhóm ...' and then instructs the agent to automatically extract secrets, persist them to environment configuration, restart the gateway, and send a live test message. Because there is no robust negative boundary or strong confirmation gate beyond a weak textual note about owner checks, this can cause unintended execution of sensitive configuration changes from conversational input, including prompt-injection or accidental activation when tokens are merely discussed.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
95% confidence
Finding

The skill instructs the agent to transmit a highly sensitive bot token to an external service by embedding it directly in the request URL path. Even though validating the token with the Zalo API is functionally necessary, this is still external transmission of a secret and is especially risky because URL-path secrets are commonly captured by process lists, proxy logs, shell history, diagnostics, and upstream telemetry. In this skill's context, the danger is elevated because the agent is also told to extract the token from chat and persist it automatically, increasing the chance of secret leakage beyond the intended destination.

Content

Scanner excerpt · SKILL.md (reported line 66)May include surrounding context.

3.1 Kiểm tra bot còn sống

bash
curl -s "https://bot-api.zaloplatforms.com/bot$ZALO_BOT_TOKEN/getMe" \
  -H "Content-Type: application/json"
  • ok: true → token hợp lệ, đi tiếp.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

Câu mẫu bắt buộc ở đây là "✅ Em đã kết nối Zalo Bot thành công.", thể hiện phong cách xưng hô/ngôn ngữ cụ thể thay vì cho người dùng lựa chọn. Theo SQP-3, việc áp đặt một ngôn ngữ hoặc locale cụ thể mà không có opt-in có thể là vi phạm chính sách ngôn ngữ.

Content

No source excerpt is available for this finding.