Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only skill for improving error messages, with broad activation wording but no hidden code, persistence, credential use, or external execution.
Installers should expect this skill to help rewrite or structure vague error messages. Review the broad implicit triggers if precise skill routing matters in your environment; otherwise the package does not show risky execution, credential, persistence, or data-access behavior.
The skill description uses very broad activation language such as general productivity, debugging, support, and implementation-help terms that can match a large range of ordinary user requests. This creates an invocation-scope problem where the skill may activate outside its intended niche, increasing the chance of inappropriate routing, prompt shadowing, or unintended influence over unrelated tasks.
The trigger keywords are highly generic terms like debugging, support, troubleshooting, and user feedback, which are common across many benign conversations and unrelated skills. Overly broad triggers can cause accidental activation and tool/skill hijacking behavior, especially in multi-skill environments where routing depends on keyword overlap.
The trigger sentence is so generic that it can match ordinary user requests unrelated to this specific skill, causing the agent to invoke the skill in unintended contexts. Over-broad activation increases prompt-surface exposure and can misroute user tasks, leading to confusing behavior, accidental disclosure of unrelated context to the skill, or unsafe downstream automation choices.
The trigger sentence is written as a broad natural-language request that overlaps with ordinary user phrasing, which can cause the skill to activate when a user is simply asking for general help with error messages. In an agent ecosystem, unintended invocation can route prompts through the wrong workflow, increasing the chance of prompt hijacking, inappropriate tool use, or confusing output without the user's explicit intent.
The invocation guidance relies on generic phrasing for a very common support/debugging task, so the routing boundary for this skill is ambiguous. Because this skill targets broadly applicable productivity scenarios, weak trigger specificity makes accidental or adversarial activation more likely and can interfere with correct skill selection.
The trigger phrases are broad enough to match many ordinary requests about productivity, support, debugging, or workflows, which can cause the skill to activate outside its intended scope. In an agent environment, over-broad invocation increases the chance of prompt-routing mistakes, unintended instruction application, or unnecessary exposure of user context to the skill.
Example trigger sentences beginning with broad language like 'Help me' are prone to accidental matches because they mirror normal user phrasing. In systems that learn or match from examples, this can further widen activation beyond the intended error-message use case and make routing less predictable.
The trigger conditions are overly broad and rely on generic terms like debugging, support, and user feedback, which can overlap with many ordinary conversations. This can cause the skill to activate unexpectedly, leading to inappropriate routing, irrelevant guidance, or overshadowing more suitable skills in ways that reduce reliability and may expose downstream workflows to unintended inputs.
The skill enables implicit invocation while describing activation in very broad terms such as work-productivity, debugging, user feedback, support, workflow, artifact, checklist, analysis, or implementation support. That makes it easier for the agent to trigger this skill in many loosely related contexts, which can cause prompt-scope creep, unintended routing, and exposure of user data or task context to a skill the user did not explicitly request.
This is a zh-CN README, but key instructional content such as the demand description, workflow explanation, and trigger phrases is presented in English. That can amount to an implicit language constraint or inconsistent locale experience without user opt-in or an explanation of intended language behavior.
No suspicious patterns detected.
