Back to skill

Security audit

Skill Crafter

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent skill-authoring helper, but it can create or modify skills and then submit/register them without a clear final user review step.

Review this skill before installing if you do not want agents to submit or re-register skills automatically. Prefer adding a mandatory confirmation step that shows the generated skill path and summary before ima_skill_create runs, and require user approval before any generated skill performs network setup such as git clone.

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
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (11)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill directs the agent to run shell commands and perform file creation/edits, but the skill frontmatter does not declare any tool scope or permission boundaries. In a system that relies on skill metadata to constrain execution, this can allow broader-than-expected filesystem or shell access and makes it harder to audit or safely sandbox the skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The activation phrases are broad enough to match ordinary discussion about skills, which can cause the agent to invoke this skill in unintended contexts. Because the skill can write files and run commands, accidental activation increases the chance of unauthorized workspace changes or unexpected registration actions based on casual conversation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The workflow instructs the agent to automatically upload/register generated skills during Phase 3 without an explicit user confirmation gate immediately before that external or persistent side effect. This is dangerous because a loosely triggered skill could publish unreviewed content, leak sensitive workflow details into a registered artifact, or create unauthorized system state changes.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
79% confidence
Finding

The file title and all checklist content are written exclusively in Chinese, and line L42 further frames English as conditional rather than offering an explicit user language choice. Under the stated policy, forcing a specific language without opt-in is a natural-language policy concern unless the locale restriction is clearly documented and justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The file presents the skill and all invocation guidance entirely in Chinese, beginning with the title and introductory description. Under the language/locale policy rule, this is a natural-language policy concern because the skill does not indicate that Chinese is optional, user-selected, or required for a justified region-specific purpose.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The guidance explicitly recommends automatically running git clone when a local path does not exist, without requiring user confirmation or discussing network, code provenance, or system modification risks. In a skill-authoring reference, this can normalize unsafe behavior and lead downstream skills to perform unprompted repository downloads and execution-adjacent setup on user systems.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This markdown guidance forces a specific language/locale for all users through its natural-language content, and it does not indicate that Chinese is optional or limited to a region-specific audience. The policy explicitly calls for flagging language or locale constraints when no user choice or justification is provided.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

该 markdown 文件从标题到全部示例均以中文固定表述工作流与触发方式,未说明是否仅适用于中文环境,也未提供用户语言选择。根据规则,若技能内容强制特定语言而无用户选择或明确且合理的地域限定,属于自然语言策略违规。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The guidance hard-codes the pattern Use when user asks to as the trigger wording, despite the rest of the document being in Chinese. This can be interpreted as forcing a specific language format rather than offering a language/locale choice or documenting a justified locale requirement.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module docstring presents usage and behavior information in Chinese only, and the rest of the script continues that pattern in printed messages. This imposes a specific language on users without any opt-in or justification that the tool is intended only for a Chinese-speaking or region-specific environment.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The generated SKILL.md template contains guidance, examples, and TODO instructions in Chinese only, which means every initialized skill inherits a forced locale by default. Because no language choice or locale justification is provided, this is a natural-language policy concern rather than a code defect.

Content

No source excerpt is available for this finding.