Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with overly broad activation wording but no hidden code, data access, persistence, or destructive behavior.
Installers should be aware that this skill may be invoked too broadly because of generic keywords and implicit invocation. It is otherwise low risk: review the output for task fit, and prefer explicit invocation when you specifically want ontology-style workflow support.
The skill description is broad enough to match many ordinary productivity, knowledge, and workflow requests, which can cause unintended activation. Over-broad routing increases the chance that this skill intercepts tasks outside its intended scope, leading to incorrect behavior, prompt-surface expansion, and possible bypass of more appropriate specialized skills.
The keyword list contains very generic terms like 'knowledge', 'graph', 'structured', 'creating', and 'bug fix' that are common across many benign requests. Such ambiguous triggers can cause widespread accidental invocation or allow a user to deliberately steer routing into this skill when it was not the best or safest match.
The trigger sentence is broad enough to match ordinary user language and may cause the skill to activate outside its intended scope. In an agent routing context, ambiguous activation can override more appropriate skills, leading to irrelevant guidance, workflow confusion, or unsafe application of generic ontology/process advice to unrelated tasks.
The trigger phrases are generic request patterns such as 'Help me' and 'I need a practical workflow', combined with broad topical language. This can cause the skill to activate for ordinary user requests that were not intended to invoke it, increasing the chance of prompt hijacking-by-overmatch, unintended workflow injection, or interference with other more appropriate skills.
The trigger phrases are broad, generic requests such as asking for a 'practical workflow' or 'help me', which can overlap with normal user language and cause the skill to activate unintentionally. In an agent environment, overbroad activation can route unrelated tasks into this skill, creating prompt-scope confusion and increasing the chance of inappropriate actions or misleading outputs.
The example trigger sentences are essentially generic help requests with no clear boundary conditions, reinforcing loose activation behavior. This makes the skill easier to invoke unintentionally and trains downstream routing or authors to treat vague phrasing as sufficient for selection.
The invocation description says to use the skill for broad topic areas or whenever the user "needs a practical workflow, artifact, checklist, analysis, or implementation support," but it does not clearly define exclusion boundaries. This ambiguity makes the skill prone to over-invocation on loosely related requests, creating misrouting risk and reducing reliability of downstream responses.
The trigger list is overly broad and includes generic terms such as "knowledge," "graph," "structured," and "creating," which are likely to match many unrelated user requests. This can cause unintended activation, routing the user into the wrong workflow and increasing the chance of irrelevant, low-quality, or unsafe task handling due to context confusion.
The default prompt is broad, vague, and phrased as a generic invocation template rather than a tightly scoped task description. This can cause the platform to activate the skill in unrelated contexts involving productivity, workflows, ontology, or implementation help, increasing the chance of unintended prompt injection exposure, user confusion, or unsafe delegation.
Enabling implicit invocation without scope constraints allows the system to call this skill automatically based on loose semantic matches. Because the skill description and prompt are themselves broad, this materially increases the attack surface for accidental or adversarial triggering in contexts the skill was not designed to handle.
This trigger remains ambiguous because it describes a broad need for a 'practical workflow' without enough constraints on the subject matter or desired output. In systems that auto-select skills from natural language, such underspecification increases misrouting risk and can cause the agent to invoke this skill for loosely related requests, reducing reliability and potentially bypassing safer or more specialized handling.
The file list explicitly labels SKILL.md as "English skill instructions" while the primary README is also in English, which suggests a default language requirement in the skill instructions. There is no statement here that the user can choose their preferred language, even though a Chinese variant exists.
This zh-CN README contains substantial English prose for core descriptive sections, effectively forcing English for parts of the user-facing experience despite being presented as the Chinese version. There is no explicit opt-in, language selection, or justification for this mixed-language behavior.
No suspicious patterns detected.
