Security audit
Security checks for vulnerabilities and agentic risk
The skill has no executable payload, but its very broad automatic activation terms could make it affect unrelated conversations.
Review before installing if you rely on precise skill routing: this skill may be invoked by ordinary words like security or bug fix, so it could steer unrelated tasks into its checklist/workflow pattern. It does not appear to install code, run commands, read secrets, or persist anything.
The manifest description is extremely broad and includes many generic task categories such as workflow, checklist, analysis, and implementation support. This can cause the skill to activate outside its intended scope, leading to prompt-routing confusion, accidental invocation in unrelated contexts, and increased exposure to adversarial or irrelevant inputs.
The keyword list contains highly generic terms like 'security', 'gate', 'every', 'must', 'pass', and 'before', which are common in ordinary conversations. This creates a high risk of unintended activation or shadowing of other skills, which can disrupt routing logic and make the agent easier to manipulate through trigger stuffing.
The trigger list includes very generic terms such as "security", "every", "must", "pass", and "before", which are likely to appear in many unrelated prompts. This can cause unintended activation of the skill, leading to prompt-routing confusion, incorrect workflow injection, or overshadowing of more appropriate skills.
The default prompt contains broad, everyday trigger terms and a generic invocation pattern that can cause the skill to activate in contexts far beyond its intended security-workflow use. Because implicit invocation is enabled, this increases the chance of unintentional routing, context pollution, or the skill being pulled into unrelated conversations where it may influence outputs or expose unnecessary functionality.
The trigger set includes very broad everyday terms such as "every" and generic security/workflow phrasing, which can cause the skill to activate in many unrelated conversations. In an agent environment, over-broad invocation increases the chance that this skill intercepts requests it was not intended to handle, leading to incorrect routing, confusing outputs, or security-review bypass if users assume the invocation was deliberate.
The trigger phrases are broad enough to match ordinary requests involving generic terms like 'security', 'workflow', 'every', 'must', or 'bug fix', which can cause the skill to activate outside its intended scope. Overbroad activation increases the chance of unintended instruction injection into unrelated conversations and may route sensitive or safety-critical tasks through a skill the user did not explicitly request.
The trigger phrases include highly generic terms such as "security," "every," "must," "before," and "bug fix," which overlap with ordinary user language and can cause the skill to activate in unrelated conversations. In an agent environment, this increases the chance of unintended routing or execution of this workflow, potentially overriding more appropriate skills and causing confusing or unsafe task handling.
The example trigger phrases are malformed, overly verbose, and do not define clear boundaries for when the skill should activate. Ambiguous examples encourage inconsistent matching behavior and make it easier for unrelated or adversarial prompts to accidentally satisfy the activation pattern.
该文件名为 SKILL.zh-CN.md 且正文整体以简体中文呈现，但文中未说明这是可选语言版本，也未向用户提供语言选择或切换方式。按规则，这构成潜在的语言/locale 策略问题，因为技能表现为默认强制特定语言而缺少用户选择。
The description says to use the skill when users ask for broad topics like work-productivity, skillscan, security, or practical workflows, but it does not clearly define boundaries or disambiguation rules. Ambiguous enablement conditions increase the chance that the skill is invoked in contexts it was not designed for, degrading reliability and potentially exposing users to irrelevant security guidance.
The trigger scope is ambiguous because it mixes category words, generic nouns, and partial natural-language sentences without defining valid activation boundaries. This can cause accidental or overly frequent activation, especially in normal productivity or security discussions, reducing reliability and potentially causing the wrong workflow to run in sensitive contexts.
This file is a Chinese README, but substantial core instructional content and trigger text remain in English, with no indication that users can choose their preferred language. That can amount to forcing a language or mixed-locale experience without explicit user opt-in.
