Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only helper for adding unit tests and improving coverage, with broad activation wording but no hidden execution, persistence, credential access, or exfiltration behavior.
This skill is reasonable to install for help with unit tests and coverage. Be aware that its automatic invocation rules are broad, so it may activate for generic testing or quality requests; explicit invocation is preferable when you want predictable routing.
The trigger sentence is broad enough to match ordinary requests about help, workflows, or tests without strong scoping, which can cause the skill to activate when the user did not intend it. In an agent system, overbroad activation can route prompts to the wrong behavior, produce irrelevant actions, and increase the chance that other safeguards or more appropriate skills are bypassed.
The trigger sentences are broad, generic help-seeking phrases that could cause the skill to activate outside narrowly intended test-coverage scenarios. In an agent ecosystem, overbroad invocation increases the chance of unintended routing, context capture, or the skill being applied to requests it was not designed to handle, which can degrade safety and predictability.
The trigger phrases are broad enough to activate on very common testing-related requests, which can cause the skill to be invoked outside its intended scope. In an agent ecosystem, overly permissive activation increases the chance of incorrect routing, prompt collisions, or unintended exposure of the skill’s instructions in contexts where a different skill should handle the task.
The skill description and invocation guidance are broad enough to match many ordinary software-testing requests, which can cause this skill to activate outside its narrowly intended context. Over-broad routing increases the chance of misselection, hidden prompt influence, or unreviewed workflow insertion into unrelated user tasks.
Generic keywords like 'testing', 'regression', and 'quality' are too common across software requests and can spuriously invoke this skill in contexts where it is not appropriate. That broad match surface makes prompt-routing less reliable and can let the skill override or interfere with more suitable specialized skills.
The description defines applicability in very broad terms, covering multiple generic software activities without clear boundaries. In an agent-routing system, ambiguous scope can cause this skill to be selected for loosely related requests, increasing the chance of irrelevant guidance, prompt collisions, or interference with more specialized skills.
The trigger keywords are broad and include common terms such as 'testing', 'regression', and 'quality', which can cause the skill to activate in contexts far beyond its intended scope. This is not code-execution dangerous by itself, but it can lead to over-invocation, misrouting, and unintended handling of user requests by this skill instead of a more appropriate one.
The default prompt embeds a very broad invocation phrase tied to generic concepts like software, unit tests, testing, and implementation support. Because these terms appear frequently in normal developer conversations, the skill may be invoked in contexts where the user did not explicitly intend to use it, increasing the risk of prompt-routing mistakes or unintended data exposure to the skill.
Implicit invocation is enabled without any visible constraints, exclusions, or confirmation step. In a broadly applicable domain like test coverage and unit testing, this creates a realistic risk that the agent is auto-selected during unrelated or only tangentially related requests, which can cause unnecessary prompt injection surface, context leakage, or unintended actions.
The trigger block lacks explicit scope boundaries, so the skill may be selected for generic software requests that mention tests or implementation help only incidentally. This ambiguity is risky because it expands the skill's activation surface, making prompt routing less predictable and easier to manipulate with loosely related phrasing.
The file states that there are English and Chinese instruction and guide files, but it does not explain how the user's language preference is selected or whether language choice is user-driven. Because language/locale policy findings apply to natural-language content, this can be interpreted as an undocumented locale behavior.
The example trigger uses everyday phrasing that is broad enough to normalize invocation for loosely related requests, reinforcing imprecise routing behavior. While not directly exploitable as code execution, it widens the set of prompts that may activate the skill and reduces invocation fidelity.
No suspicious patterns detected.
