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 run code, request credentials, persist state, or perform hidden actions.
Installers should expect this skill to activate broadly for productivity or workflow-like requests. It appears safe from a security perspective, but users who run many skills may want the publisher to narrow the trigger terms or disable implicit invocation to avoid irrelevant workflow guidance.
The trigger phrases are broad enough to match ordinary productivity requests, which can cause the skill to activate outside its intended scope. In an agent ecosystem, overbroad activation can route unrelated user tasks through this workflow, creating prompt-surface expansion, unexpected behavior, and possible interference with safer or more appropriate skills.
The manifest description uses expansive invocation guidance like 'use when a user asks for work-productivity... transform, task, or needs a practical workflow', which is broad enough to match many ordinary requests outside the skill's intended domain. In a multi-skill environment, this can cause over-selection of the skill, reducing routing precision and increasing the chance of irrelevant or unsafe workflow generation in the wrong context.
The trigger list is overly broad and includes common words such as 'task', 'needs', and 'partners', which can cause the skill to activate for many unrelated prompts. This creates prompt-routing risk: users may be steered into this workflow unexpectedly, leading to inappropriate automation, confusing responses, or accidental precedence over more suitable skills.
The trigger sentence is an extremely broad, everyday phrase ('Help me ...'), which can cause the skill to activate for many unrelated user requests. In an agent environment, overbroad activation increases the chance of unintended routing, context capture, and execution of this workflow when a more appropriate or safer skill should handle the task.
The trigger phrases are broad enough to match generic productivity or workflow requests, which can cause the skill to activate outside its intended scope. In an agent environment, overbroad activation can misroute user intent, override more appropriate skills, and increase the chance of unintended instruction injection through loosely matched contexts.
This README is presented entirely in Chinese, while the described users include broad "AI-agent users, skill authors, maintainers, and teams" rather than a clearly China-specific audience. There is no visible statement that the user can choose language, nor a documented justification that the skill is restricted to a Chinese locale.
The skill description defines activation conditions with a wide set of generic topics and outputs, including broad requests for practical workflows, artifacts, checklists, analysis, or implementation support. This unclear boundary makes the skill eligible for many adjacent tasks, which can lead to accidental invocation, prompt-routing errors, and unintended influence over unrelated user requests.
The trigger list includes very broad everyday terms such as "task", "needs", and "bug fix", which can cause the skill to activate in many unrelated conversations. Over-broad activation increases the chance that the agent will inject irrelevant workflow guidance into contexts where it was not requested, reducing reliability and potentially interfering with other more appropriate skills.
The manifest description and default prompt are overly broad and vague, while implicit invocation is enabled. This can cause the skill to trigger in contexts beyond its intended scope, leading to unexpected handling of user requests, confused-deputy behavior, or accidental application of workflow guidance where it was not explicitly requested.
This trigger sentence is ambiguously scoped ('I need a practical workflow for ...') and can match a wide range of normal requests that are not specifically about this skill's intended domain. Such ambiguity makes accidental invocation more likely, which is dangerous in multi-skill systems because it can lead to incorrect assistance, unnecessary data exposure to the skill, or bypass of more suitable workflows.
The file list presents separate English and Chinese documentation variants, but this README does not state that the user can choose their preferred language or locale. Under the language policy, a skill should explicitly offer language choice rather than implicitly fixing behavior by locale-specific variants alone.
