Back to skill

Security audit

数据映射与队列分析 (Agentic AI 科研平台)

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small, disclosed cohort-analysis helper that reads one specified demo data file and sends status-only updates to localhost, with no persistence or destructive behavior.

Install only if you expect a Chinese-language cohort-feasibility helper that can read the specified local mock cohort file and report run status to a local service on port 5001. Confirm that the localhost reporting endpoint is part of your environment and that aggregate cohort counts are acceptable to send there.

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 (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest description says the skill triggers whenever the user needs data mapping or cohort feasibility analysis, but it does not define clear boundaries or exclusions. This broad phrasing could cause unintended invocation for general analytical requests that mention variables, cohorts, or feasibility without specifically requesting this skill.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The natural-language description and user-facing messaging are entirely in Chinese, with no indication that language choice is optional or based on user preference. This can violate language/locale policy when the skill implicitly forces a specific language without explicit opt-in or documented regional justification.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The example phrase “帮我看看院内有哪些可用变量” is a common exploratory request and is not uniquely tied to this specific skill. Without additional constraints, such a generic phrase could collide with other data-discovery or analytics skills.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
96% confidence
Finding

The skill instructs the agent to send execution status to an HTTP endpoint on localhost, which is an external transmission channel outside the immediate user response path. Even though the sample payload contains only status metadata, this pattern can normalize silent side-channel reporting and may expose workflow details or become a vehicle for broader data leakage if later expanded; the use of unauthenticated plain HTTP also weakens integrity and confidentiality.

Content

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

1. 上报开始

bash
curl -s -X POST http://localhost:5001/api/report \
  -H "Content-Type: application/json" \
  -d '{"skill":"cohort-analyst","display_name":"数据映射与队列分析","status":"running","message":"正在连接院内数据字典,匹配变量与评估缺失率..."}'