Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only helper for unit-test and coverage work, with no executable code, persistence, credential handling, or hidden behavior found.
Installers should be aware that the skill may activate for broad testing or quality-related requests because implicit invocation is enabled, but its contents are limited to local unit-test and coverage assistance and do not introduce executable code or privileged behavior.
The trigger sentence is broad enough to match ordinary user phrasing, which can cause the skill to activate outside its intended scope. Over-broad activation can route unrelated requests into this skill, leading to inappropriate guidance, prompt hijacking opportunities through misrouting, or reduced reliability of skill selection.
The invocation wording uses loose natural-language phrasing without clear scope, increasing the chance of accidental or ambiguous activation. In agent environments, ambiguous routing can expose users to the wrong workflow or allow an attacker to intentionally craft prompts that invoke this skill when another tool or policy should apply.
The trigger phrases are broad enough to match common requests about testing, coverage, workflows, or implementation help, which raises the chance of accidental or over-frequent activation. In an agent ecosystem, unintended invocation can route user requests into a skill that shapes behavior or outputs without explicit user intent, creating reliability and prompt-scope risk even if the content is not overtly malicious.
The skill description is broad enough to match many ordinary software-help requests, which can cause the skill to activate outside its intended testing and coverage scope. Over-broad routing increases the chance of inappropriate instruction injection into unrelated tasks, reducing control over which skill handles a request.
Generic trigger keywords like 'software-and-data', 'testing', and 'quality' are too common across many benign prompts, making accidental invocation likely. In a multi-skill environment this can misroute user requests, override better-scoped skills, or expose the user to irrelevant workflows and instructions.
The trigger description is overly broad and mixes a wide set of generic software topics such as software-and-data, testing, regression, and implementation support. In an agentic environment, this can cause the skill to activate for requests outside its intended scope, leading to inappropriate routing, over-collection of user context, or unreviewed test/code suggestions being applied where they do not fit.
The keyword list contains generic terms like testing, regression, and quality without negative examples or disambiguation. This increases the chance of accidental invocation on unrelated prompts, which can misroute users, crowd out more appropriate skills, and produce unsafe or low-relevance automated guidance in contexts not actually about unit-test coverage.
The default prompt uses a very broad natural-language invocation phrase ('help me' plus common testing-related terms) while implicit invocation is enabled. This increases the chance the skill will be triggered during ordinary conversation about tests or coverage, causing unintended routing, prompt interference, or inappropriate access to the skill when the user did not explicitly request it.
The trigger phrase is broad enough to match ordinary requests about testing or coverage that may not specifically require this skill, which can cause inappropriate auto-invocation or routing. In an agent system, overly generic triggers increase the attack surface for prompt confusion, unintended tool use, and misapplication of the skill in contexts where its assumptions do not hold.
This file is explicitly a Chinese-language README and references parallel English and Chinese documentation, but it does not state that the user can choose their preferred language or that Chinese output is opt-in. That can create a language/locale policy concern if the skill defaults users into a specific locale without consent.
The example trigger sentences use vague, natural phrasing that does not clearly distinguish this skill from general assistance requests. While less severe than broad metadata fields, these examples can still normalize over-activation and lead to ambiguous routing behavior.
No suspicious patterns detected.
