ClawHub
Back to skill

Security audit

Work Productivity Self Improving Workflow Helper

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only workflow helper with overly broad activation wording but no code execution, persistence, credential use, or hidden data access.

Installers should be aware that this skill may activate too easily because of broad keywords and implicit invocation; use it for explicit workflow or agent-reliability planning tasks, and prefer more specialized skills for unrelated requests.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (12)

Vague Triggers

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger phrases are broad and include generic wording such as 'help me' and 'I need a practical workflow', which can cause the skill to activate in many ordinary conversations unrelated to this specific capability. In an agent ecosystem, unintended invocation can override user intent, route tasks to the wrong workflow, and increase the chance that proactive or self-improving behavior is applied without clear consent.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest description defines a catch-all invocation condition using broad concepts like "work-productivity" and "practical workflow, artifact, checklist, analysis, or implementation support," which describes a huge portion of normal assistant activity. This creates routing ambiguity and makes accidental or excessive invocation likely, reducing predictability and potentially causing this skill to override more specialized skills.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger keywords are extremely broad and include common words like "self," "improving," and "proactive," which can cause the skill to activate during ordinary conversations unrelated to this skill’s purpose. In an agent environment, overbroad activation can hijack routing, inject irrelevant workflow behavior, and increase the chance that users are guided by this skill when another, more appropriate skill should handle the task.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

技能描述中的启用条件使用了宽泛枚举，如“work-productivity, self-improving-proactive-agent, self, improving, proactive”，边界不清，几乎可覆盖大量一般性生产力或自我改进请求。这会让路由器或代理在缺乏明确用户意图时加载该技能，增加上下文污染、错误委派和非预期执行路径的风险。

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

触发关键词包含“self”、“improving”、“proactive”等高频日常词汇，且示例触发句也非常宽泛，容易在与该技能无关的普通对话中被误触发。误触发后，代理可能偏离用户原始任务，插入不必要的流程、分析或自动化建议，造成行为扩大和结果污染。

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger sentences use generic, natural-language phrasing such as 'Help me' and 'I need a practical workflow', which overlaps heavily with ordinary user requests. This can cause the skill to activate unintentionally in unrelated contexts, leading to prompt routing errors, confusing responses, or the skill being applied to requests the user did not intend for it.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger examples are extremely generic and can cause the skill to activate for broad, ambiguous requests unrelated to the intended workflow. In an agentic context, over-broad activation increases the chance of unintended instruction capture, workflow hijacking, or interference with other safer/more appropriate skills.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The example trigger sentences reinforce vague activation patterns by pairing generic lead-ins like "Help me" and "I need a practical workflow" with loosely bounded problem statements. These examples train or encourage matching behavior that is too permissive, increasing accidental activation and making the skill harder to safely compose with other skills.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
80% confidence
Finding

该技能文件为 SKILL.zh-CN.md，内容整体以简体中文（中国地区）呈现，但文中未说明这是用户可选语言版本，也未提供语言/区域选择或切换指引。按规则，若技能强制特定语言/locale 且无用户 opt-in，应视为自然语言策略风险。

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill’s invocation guidance is overly broad and paired with allow_implicit_invocation: true, which increases the chance the agent will invoke this skill in situations only loosely related to the user’s request. That can cause unintended autonomy, prompt-scope creep, and misuse of a self-improving/proactive workflow helper in contexts where the user did not clearly consent to that behavior.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
83% confidence
Finding

The file advertises both English and Chinese instruction and guide files, but does not state how the user's language preference is selected. This can create a locale-policy ambiguity if the skill defaults to a language without explicit user choice or documented justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
76% confidence
Finding

The file is presented as a Chinese README, but substantial user-facing instructional content and trigger phrases remain in English. This can impose a language choice on users without explicitly offering alternatives or documenting the language expectation.

Content

No source excerpt is available for this finding.