Security audit
Security checks for vulnerabilities and agentic risk
This non-executable workflow helper is not malicious, but its implicit activation is too broad and could affect unrelated conversations.
Review this skill before installing if you rely on implicit skill routing. It should be limited to explicit PollyReach-style workflow or productivity-hardening requests, and the generic trigger terms should be narrowed or implicit invocation disabled.
The trigger phrases are broad enough to match ordinary user requests such as asking for help, practical workflows, or bug fixes, which can cause this skill to activate outside its intended scope. In an agent ecosystem, overbroad activation can route unrelated conversations into this skill, leading to unintended instruction injection exposure, user confusion, or inappropriate task handling.
The description uses an ambiguous activation condition: "Use when a user asks for work-productivity, pollyreach, gives, every, phone, or needs a practical workflow..." This broad phrasing makes the skill eligible for many unrelated contexts and increases the chance of accidental invocation, which is dangerous because a misrouted skill can steer the agent away from the user's actual intent and apply irrelevant workflows or instructions.
The trigger keyword list is excessively broad and includes generic terms like "every," "phone," "number," "ability," "things," and "done" that commonly appear in ordinary conversations. This creates a prompt-routing vulnerability where the skill may activate unintentionally for unrelated user requests, causing scope confusion, inappropriate instruction injection into unrelated tasks, and degraded safety controls through accidental skill selection.
The trigger list includes extremely common terms such as "every", "phone", "number", "ability", "things", and "done", which can cause the skill to activate for many unrelated user requests. This creates routing confusion and may surface inappropriate workflow guidance in contexts where the user did not intend to invoke this skill, reducing reliability and potentially interfering with safer or more relevant skills.
The trigger guidance includes highly generic language that can match ordinary user requests unrelated to this skill’s intended scope, increasing the chance the skill is invoked inappropriately. Over-broad activation can cause routing errors, irrelevant assistance, and unexpected execution of workflow logic in contexts where the user did not ask for this capability.
This file presents the user-facing skill documentation entirely in Chinese, while describing a general productivity workflow rather than a clearly China-specific or Chinese-only compliance use case. There is no visible statement in this file that the user can choose another language or that the locale restriction is intentional and justified.
The trigger phrases are extremely broad and include generic terms like 'phone', 'number', 'ability', 'things', and 'done', which can cause the skill to activate on many unrelated user requests. In an agent ecosystem, this creates prompt-routing confusion and can hijack normal workflows, causing unintended execution of this skill instead of a more appropriate one.
The example trigger sentences rely on generic phrases like "Help me" and "I need a practical workflow," with only loosely attached domain text. These examples reinforce overbroad matching behavior and may cause downstream systems or authors to believe vague language is sufficient for activation, further increasing false activations and task hijacking risk.
The invocation description says to use the skill for broad categories like work-productivity or whenever a user needs practical workflow, checklist, analysis, or implementation support for the requirement, which is overly expansive. This ambiguity increases the chance of unintended activation and misrouting, especially because the skill appears to be framed around a narrow PollyReach-style workflow use case but describes applicability in much broader terms.
The skill enables implicit invocation while using a very broad default prompt and description terms such as general productivity/help language, which creates weak trigger boundaries. This can cause the skill to activate in unrelated conversations, unexpectedly influencing agent behavior or pulling in workflows the user did not explicitly request, increasing the risk of prompt-scope confusion and unsafe action routing.
The invocation guidance lacks precise boundaries and mixes specific terms with broad phrases, which can cause accidental triggering across unrelated work-productivity conversations. In an agent environment, this kind of overmatching can misroute requests, override better-matched skills, and reduce trust in tool behavior.
No suspicious patterns detected.
