Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper with overly broad activation wording, but no evidence of hidden execution, data access, persistence, or destructive behavior.
Before installing, review whether you want a skill with broad terms like search, engine, and bug fix to invoke implicitly. It appears safe in behavior, but its activation wording may make it show up for unrelated requests until narrowed.
The trigger keyword list includes highly generic terms such as "multi", "search", "engine", and "bug fix", which are common across many unrelated requests. In agent routing systems, this can cause unintended activation, leading the wrong skill to intercept prompts, produce irrelevant guidance, or displace more appropriate skills.
The default prompt is phrased as a broad, natural-language invocation trigger tied to common terms like work productivity, search, and workflow help, which increases the chance of unintended or implicit activation. Because the policy also allows implicit invocation, unrelated user requests could route into this skill unexpectedly, expanding attack surface and enabling prompt/context injection through overbroad matching.
The trigger sentence is so broad and conversational that it could cause the skill to activate on unrelated user requests containing common terms like "help me" plus generic workflow language. In an agent-routing context, overly permissive activation increases the chance of unintended invocation, misrouting, and execution of the wrong workflow, which can degrade reliability and potentially expose downstream tools or actions to inappropriate contexts.
The trigger phrases are extremely broad and partially generic, which can cause the skill to activate in unrelated contexts. In an agent ecosystem, over-broad invocation can route user requests into the wrong workflow, leading to inappropriate guidance, unintended processing, or reduced reliability of downstream safety controls.
The trigger phrases are broad, generic, and include common terms like 'multi', 'search', 'engine', and vague help-request templates, which can cause the skill to activate outside its intended scope. In an agent environment, unintended invocation can route user requests into the wrong workflow, producing irrelevant actions, unsafe automation chaining, or bypass of more appropriate specialized skills.
The manifest description uses broad matching language like "Use when a user asks for work-productivity... or needs a practical workflow, artifact, checklist, analysis, or implementation support," which describes a very wide range of normal requests. In a skill-selection pipeline, this increases the chance of overmatching and accidental routing, reducing reliability and potentially exposing users to inappropriate or lower-quality task handling.
The example trigger phrases are written in broad, everyday wording and effectively suggest activation from loosely related user language. This reinforces overbroad routing behavior and may train maintainers or automated systems to accept ambiguous activations, causing prompt interception by this skill outside its intended scope.
The skill description says it should be used whenever a user mentions broad categories like work-productivity, multi, search, or engine, without defining clear boundaries or exclusions. This makes the skill prone to over-triggering in benign unrelated contexts, which can degrade agent reliability and allow the wrong workflow to influence task handling.
The trigger list includes very broad generic terms such as “multi”, “search”, “engine”, and “global”, which can match many unrelated user requests and cause this skill to activate unexpectedly. Over-broad activation increases the chance of workflow hijacking, irrelevant instruction injection into conversations, and accidental substitution for a more appropriate skill.
The activation guidance lists generic keywords such as "multi," "search," and "engine," which are individually common across many benign requests and therefore weak selectors for this specific skill. This makes accidental activation more likely, causing incorrect task routing and reducing trust in the agent system; if the skill performs actions or generates artifacts, that misrouting can have broader operational effects.
The file descriptions explicitly label separate English and Chinese instruction files, but the README does not state that users can choose their preferred language or how language selection occurs. That can create an implied locale policy constraint without documented opt-in.
The file presents user-facing instructions in Chinese and identifies the target audience broadly as AI-agent users, skill authors, maintainers, and teams, but it does not indicate that Chinese is optional or region-specific. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy concern.
Several evidence items are presented in Chinese while the rest of the document is in English, but the file does not explain any language-selection policy or offer a user choice. This can create an implicit locale inconsistency in skill materials without documented opt-in or justification.
