Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with overly broad activation wording, but it does not install code, request credentials, or perform actions on its own.
Before installing, be aware that this skill may activate for broad Google, workspace, CLI, or bug-fix requests. Install it if you want a general workflow helper for Gog/Google Workspace style productivity guidance; consider narrowing or disabling implicit invocation if precise routing matters.
The manifest description uses very broad activation language such as common domains and generic request types, which can cause the skill to match many unrelated user prompts. In an agent routing system, overbroad matching can trigger this skill unexpectedly, leading to prompt hijacking of normal workflows, reduced reliability, and possible exposure of user context to an unnecessary skill.
The trigger keyword list includes vague, high-frequency terms like 'google', 'workspace', 'cli', and 'bug fix' without contextual qualifiers. These terms are common across many benign requests, so the skill may activate far outside its intended domain and interfere with routing or capture tasks better handled by another skill.
The trigger list is very broad and includes generic terms such as "google", "workspace", "cli", and "bug fix", which are common across many unrelated user requests. This can cause the skill to activate in contexts it was not intended for, leading to incorrect routing, accidental override of more specific skills, and potentially unsafe or irrelevant assistance being surfaced.
The trigger sentence uses very broad, everyday wording and also includes generic terms like 'help me' and 'practical workflow,' which can cause the skill to activate outside its intended scope. In an agent system, overbroad activation can route unrelated user requests into this skill, creating confused-deputy behavior, poor task selection, and accidental execution of workflows not meant for the current context.
The trigger examples are so broad and repetitive that they could cause the skill to activate for loosely related requests involving Google, workspace, or generic workflow help. In an agent environment, ambiguous activation increases the chance of unintended use, context confusion, or the wrong workflow being applied to sensitive productivity tasks such as Gmail, Calendar, or Drive operations.
The trigger phrases are broad, natural-language prompts that can match ordinary user requests unrelated to this specific skill. That increases the chance of accidental invocation, causing the agent to apply workflow guidance or operational behavior in contexts the user did not explicitly intend, which can lead to confusion, mis-execution, or unintended access to connected productivity tooling.
The example trigger sentences rely on generic lead-ins like 'Help me' and 'I need', which are extremely common in ordinary user requests. Although examples are not executable code, they can reinforce lax activation patterns in downstream tooling or human authors, increasing accidental invocation and weakening skill-selection precision.
This zh-CN skill file contains core instructional content and examples in English, including the target user description and validated-demand text, without indicating that the user can choose their preferred language. That creates a locale/language policy concern because the file implicitly forces mixed-language interaction rather than clearly offering opt-in language selection.
The default activation prompt is broad and matches generic help-seeking language such as asking for practical help, workflows, checklists, analysis, or implementation support. Because implicit invocation is enabled, this can cause the skill to trigger in situations far beyond its intended scope, increasing the chance of unintended tool behavior, context capture, or prompt-surface expansion in ordinary conversations.
The activation guidance is ambiguous and does not clearly define when the skill should and should not be selected. Because the skill metadata already spans broad terms like work-productivity, google, workspace, and cli, the weakly constrained trigger section increases the chance of unintended invocation and misrouting of user tasks.
The file lists only English and Chinese instruction and guide files, which may imply a fixed language set without stating that users can choose their preferred language. Because locale policy issues concern forcing a specific language without opt-in, this is a mild documentation-level concern.
This file is presented as a Chinese README, but core descriptive and usage content is written in English, including the main requirement text and trigger phrasing. That can amount to a language-policy mismatch because the skill appears to force English content in a locale-specific document without user opt-in or documented justification.
