Security audit

Skill Forge 2026 03 21

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed skill-building toolkit with local testing, security-checking, token analysis, templates, and Git helper scripts, with no evidence of hidden exfiltration or unrelated destructive behavior.

Install this if you want a local toolkit for creating and checking OpenClaw skills. Treat the bundled Git helpers as real repository-changing commands, review the target path before running tests or scans, and require explicit confirmation for any generated file-delete, move, rename, commit, branch, or publish workflow.

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (7)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The trigger phrases are very broad, covering generic requests like creating, optimizing, auditing, testing, publishing, and security review. Overbroad triggers can cause the skill to activate in many normal conversations and then steer the model into using its file/shell workflow unnecessarily, expanding the attack surface and increasing the chance of unintended actions.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger phrases are very broad generic requests such as helping call an API, fetch service data, or send data to an API. In a skill-routing system, this can cause the skill to activate for many ordinary user requests and potentially route sensitive or unintended work into a template that encourages external API interaction, increasing the chance of data exfiltration, unintended network access, or confused-deputy behavior.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The trigger phrases are broad, natural-language requests such as '帮我自动化这个流程' and '生成每日报告', which can overlap with ordinary conversation and cause the skill to activate unintentionally. In an automation-oriented skill, accidental activation is more dangerous because it may lead to workflow generation, scheduling, or monitoring actions being initiated in the wrong context.

Vague Triggers

Medium

Confidence: 85% confidence
Finding: The trigger phrases are broad, generic requests such as querying data, analyzing data, and generating visualizations, which can match many ordinary user prompts and cause the skill to activate unexpectedly. In a skill-building toolkit context, overbroad routing increases the chance that unrelated conversations are pulled into this skill, which can lead to unintended data-handling behavior or inappropriate tool invocation.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The trigger phrases are broad, everyday requests such as converting, batching, compressing, or extracting file contents, which can cause the skill to activate in situations the user did not clearly intend. Because this skill is designed for file operations, unintended invocation could lead to unexpected processing of user files or sensitive content exposure through subsequent steps.

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The skill advertises capabilities including writing, renaming, moving, and deleting files without any visible warning, safeguards, or confirmation model for potentially destructive actions. In a file-processing context, this materially increases the risk of accidental data loss, overwriting important files, or bulk destructive changes if the skill is invoked improperly or used on the wrong path.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The trigger phrases are broad everyday requests such as '帮我提交代码', '查看仓库状态', and especially '撤销' or '恢复', which can plausibly appear in normal conversation and accidentally invoke the skill. Because this skill performs Git operations that may change repository state, unintended activation could lead to unwanted commits, branch switches, or rollback-related actions, increasing operational risk even if the author likely intended convenience rather than abuse.

Static analysis

No suspicious patterns detected.