Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only unit-test coverage helper with no executable code, persistence, credential handling, or hidden behavior, though its automatic invocation triggers are broader than ideal.
Safe to install for help with unit tests and coverage. Be aware it may activate on broad words like testing or quality; explicitly invoke it for coverage work or disable implicit invocation if you want tighter routing.
The trigger sentence is so broad and natural-language-like that it can match ordinary user phrasing and cause the skill to activate outside its intended scope. In an agent system, over-broad routing can redirect benign requests into unnecessary code/test-generation workflows, increasing the chance of inappropriate actions, prompt-scope confusion, or misuse of repository context.
The trigger sentences are templated so broadly that they can match ordinary requests for help with tests or practical workflows, increasing the chance this skill is invoked when a more specific or safer routing decision should be made. Overbroad activation can cause unintended context capture, tool use, or response shaping across many unrelated software requests, which is a real security and reliability concern in agent dispatch systems.
The file is a Chinese-language README and does not indicate any user choice, fallback, or opt-in regarding language selection. For a general-purpose skill, presenting instructions in a fixed locale without documenting language choice can violate language/locale policy expectations.
The trigger phrases are very generic and include broad terms like 'testing', 'regression', and direct invocation text that can match many ordinary software requests without meaningful constraints. This can cause the skill to activate in contexts where it is not the best fit, potentially overriding more appropriate safety, domain, or workflow guidance and increasing the chance of unintended behavior in agent routing.
The trigger keywords are broad enough to match many ordinary software requests, which can cause this skill to activate outside its intended scope. Over-broad activation increases the chance of inappropriate routing, prompt hijacking exposure, or the skill being used in contexts where its assumptions and workflows do not fit.
The example trigger phrases use everyday language patterns that are so generic they may capture unrelated user requests. This makes accidental invocation more likely and can broaden the attack surface by pulling the skill into conversations where an attacker can exploit its authority or where it displaces a more appropriate skill.
该技能文件整体以中文呈现，且未声明这是可选语言版本、也未说明用户可选择其他语言输出。按规则，若技能强制特定语言而无用户选择或明确合理依据，属于自然语言策略违规。
The description says to use the skill for several broad topics but does not clearly define where its responsibility ends or when it should not be used. This ambiguity can lead to accidental invocation for general software/data requests, producing mis-scoped actions or overshadowing more appropriate specialized skills.
The trigger keywords include broad, generic terms such as "testing" and "quality", which can cause the skill to activate in contexts far beyond unit-test coverage assistance. Over-broad invocation increases the chance of unintended routing, irrelevant guidance, or interference with other skills handling unrelated quality topics.
The default prompt uses very broad natural-language triggers like 'software-and-data', 'unit tests', 'test coverage', 'testing', 'analysis', and 'implementation support', which can match many ordinary requests outside a tightly bounded scope. In combination with automatic routing, this increases the chance the skill is invoked in contexts the user did not explicitly intend, potentially causing prompt-scope confusion, misrouting, or unnecessary exposure of repository context to the skill.
Enabling implicit invocation without precise activation constraints allows the system to call this skill based on ambiguous user phrasing rather than explicit intent. Because this skill is framed broadly around testing and implementation support, ambiguous auto-invocation can cause the agent to apply the skill in unrelated development tasks, increasing the risk of unintended behavior, scope expansion, or user confusion.
The invocation guidance lacks precise boundaries for when the skill should and should not run, which makes accidental or overly aggressive activation likely. Because this skill can produce implementation workflows and code-related artifacts, ambiguous triggering increases the risk of unintended execution paths, noisy dispatch, and actions taken on insufficiently relevant requests.
