Back to skill

Security audit

Attachment Inject

Security checks for vulnerabilities and agentic risk

Overview

This skill is not malicious, but it asks to manage dynamic agent and skill context with broad read/write behavior that deserves review before installation.

Install this only if you want a skill that can influence how agents and skills are discovered, summarized, and injected into context. Review the workspace paths it may read or update, and prefer explicit user approval or allowlists before generating registries or modifying AGENTS.md.

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
89% confidence
Finding

The trigger examples are intentionally broad (e.g. matching generic mentions like '调试' or '调度Agent') and provide no authorization, context, or task-scoping guardrails before reading additional agent/skill files. In a system that can read local skill definitions dynamically, this can cause unintended capability activation or prompt-surface expansion from casual user phrasing, increasing the chance of privilege creep and unsafe chaining.

Content

No source excerpt is available for this finding.

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
83% confidence
Finding

The skill explicitly describes reading agents/*.md and a specific skills/.../SKILL.md on demand, which reveals the internal file layout and encourages runtime enumeration of available agents/skills. In agentic systems, this kind of discoverability can help an attacker probe capabilities, trigger unintended tool use, or exfiltrate metadata about internal prompt assets even if full contents are not preloaded.

Content

Scanner excerpt · SKILL.md (reported line 23)May include surrounding context.

不预注入内容,而是在需要时读取:

text
用户提到"调度Agent" → 读 agents/*.md → 选择合适的Agent
用户提到"调试" → 读 skills/systematic-debugging/SKILL.md → 激活技能

模式2: 注册表文件(轻量注入)

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

Natural-language content in the skill file is primarily in Chinese, and the document does not indicate that this language choice is optional, user-selected, or required for a specific regional context. This can conflict with language/locale policies that require user choice or explicit justification.

Content

No source excerpt is available for this finding.