Security audit
Security checks for vulnerabilities and agentic risk
The skill appears non-destructive, but it can auto-activate for very broad ordinary writing, editing, text, and bug-fix requests.
Review this skill before installing if your environment relies on implicit skill routing. Its content is not malicious, but the activation terms are broad enough that it may be selected for unrelated writing, editing, text cleanup, or bug-fix requests unless invocation is narrowed or made explicit.
The invocation description uses expansive, catch-all language such as 'practical workflow, artifact, checklist, analysis, or implementation support' for a wide range of user asks. Without clear scope restrictions, the skill may be selected for many unrelated tasks, increasing the chance of unintended behavior, incorrect tool choice, and policy bypass through overbroad delegation.
The trigger list is excessively broad, including generic terms like 'writing', 'text', 'editing', 'reviewing', and 'bug fix' that commonly appear in unrelated requests. This can cause the skill to activate outside its intended domain, leading to prompt hijacking of normal tasks, incorrect routing, or unreviewed workflow injection into benign user interactions.
The trigger list includes very broad, everyday terms such as "remove", "text", "editing", "reviewing", and "bug fix", which can cause the skill to activate in many unrelated conversations. Overbroad invocation increases the chance of unintended workflow execution, context hijacking, or the model prioritizing this skill when another, more appropriate skill should handle the request.
The trigger sentence is broad and resembles normal user language, which can cause the skill to activate when a user did not explicitly intend to invoke it. In an agent environment, unintended invocation can route tasks to the wrong workflow, causing confusing outputs, prompt-scope drift, or accidental execution of inappropriate instructions for unrelated requests.
The trigger phrases are broad, generic, and include common help-seeking language such as asking for a practical workflow or help fixing bugs. This can cause the skill to activate on unrelated user requests, leading to unintended routing, prompt interference, or accidental application of the skill in contexts the user did not intend.
The trigger phrases are broad, generic, and partially templated, which increases the chance that normal user requests will invoke this skill unintentionally. In an agent ecosystem, accidental invocation can route tasks to the wrong workflow, causing inappropriate text transformation, unexpected actions, or confusion about which skill is operating.
The example trigger phrases are vague and effectively restate generic productivity needs rather than demonstrating clear activation conditions. Ambiguous examples train downstream selection systems to invoke the skill too often, which increases misrouting risk and makes security review of activation pathways less reliable.
The skill is presented only as a zh-CN document without any language selection guidance or fallback behavior. This can cause misunderstanding of scope, misuse by non-Chinese users or maintainers, and operational errors if the wrong audience invokes or edits the skill without fully understanding its instructions.
The description says to use the skill when users mention broad categories like work-productivity, humanizer, remove, signs, or generated, or whenever they need practical workflow or implementation support. This vague boundary makes it hard to distinguish intended use from ordinary requests, increasing accidental triggering and inappropriate routing of user tasks.
The manifest prompt text is vague and overly broad, describing a wide range of loosely related tasks rather than a clearly bounded capability. Because implicit invocation is enabled, this ambiguity can cause the skill to be selected in contexts the user did not intend, increasing the risk of prompt-scope overreach, unintended workflow changes, or unsafe assistance in adjacent domains.
The trigger description defines activation using a wide set of generic terms like work-productivity, writing, text, editing, reviewing, and bug fix, without clear boundaries for when this skill is appropriate. That increases the chance of accidental selection over more suitable skills, which can lead to misrouting, reduced reliability, and unsafe context mixing if the skill is applied outside its intended domain.
The evidence list includes Chinese-language entries in an otherwise English README, but there is no note explaining language expectations or offering a user-facing language choice. This can violate language/locale policy expectations when a skill appears to assume multilingual handling without explicit opt-in or justification.
The file is a Chinese README, but a core descriptive instruction is written in English without any indication that the user can choose language or that English is required for a region-specific purpose. This can create a language-policy concern because the documentation implies a fixed language presentation without explicit opt-in.
