Security audit
Security checks for vulnerabilities and agentic risk
This is a low-impact workflow-helper skill with no executable code, but its activation phrases are overly broad and may trigger when it is not relevant.
Install only if you want a general PollyReach-style workflow helper. Be aware that the current trigger metadata is loose and may cause the skill to appear for unrelated productivity or phone/number-related requests; the publisher should narrow those triggers and implicit invocation criteria.
The trigger phrases are so broad and generic that they can match many ordinary user requests unrelated to this skill’s intended scope. That increases the chance of accidental invocation, skill overreach, and misrouting users into workflows they did not intend, which is a real safety and reliability issue for agent behavior.
The trigger list includes very generic terms like 'every', 'phone', 'number', 'ability', 'things', and 'done', which are likely to appear in many unrelated user requests. This can cause unintended skill activation, leading the agent to apply the wrong workflow or expose users to irrelevant automation behavior in contexts where this skill was not intended.
The trigger section includes extremely broad, common terms such as "every", "phone", "number", "ability", "things", and "done", which can cause the skill to activate in many unrelated conversations. Over-broad activation increases the chance of unintended routing, prompt/context injection into the wrong workflow, and user confusion, especially because this skill is framed as a general workflow helper rather than being tightly scoped.
The trigger sentences and keyword list are broad enough to match many ordinary requests unrelated to this specific skill, increasing the chance the agent invokes it in the wrong context. That can cause skill hijacking or misrouting, where users receive irrelevant workflow behavior instead of the most appropriate tool, degrading reliability and potentially bypassing safer or more specific skills.
The trigger phrases are generic and loosely scoped, using common terms like 'help me', 'practical workflow', and broad productivity-related keywords. This can cause unintended activation in unrelated conversations, leading the agent to apply the skill outside its intended context and increasing the chance of unsafe, confusing, or inappropriate workflow generation.
This file is presented as the Chinese README, but substantial user-facing content such as the demand description is written in English. That can effectively force a language choice on users of the localized document without opt-in or explanation, which is a natural-language policy concern.
The example triggers use broad starter phrases like 'Help me' and 'I need a practical workflow' with only weak domain anchoring, which encourages overmatching by activation systems or downstream authors copying the pattern. In practice, this increases accidental invocation and can cause the assistant to steer benign, unrelated tasks into this skill's workflow.
The top-level description says to use the skill when users mention broad terms like work-productivity, pollyreach, gives, every, or phone, but it does not clearly define boundaries, exclusions, or the exact user intent required. This ambiguous enablement scope makes accidental invocation more likely and can route unrelated tasks into this skill, reducing reliability and potentially exposing irrelevant context to the wrong workflow.
The default prompt embeds a very broad natural-language trigger phrase tied to common workplace terms, which increases the chance of accidental or context-insensitive invocation. In an agent environment, this can cause the skill to activate when users discuss ordinary productivity topics, potentially routing requests through unintended instructions or behaviors.
Enabling implicit invocation without constraints allows the platform to auto-select this skill based on loosely related user text, which raises the risk of unintended activation and prompt-routing errors. Combined with the broad productivity-oriented scope of the skill, this makes accidental invocation more likely and can expose users to actions or guidance they did not intentionally request.
