Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only workflow helper, but its activation wording is broad enough that users should watch for accidental invocation.
Install only if you want a broad SkillScan-style workflow assistant. Because implicit invocation is enabled and the trigger language is generic, review activation behavior and prefer explicit invocation if your environment supports it.
The skill description says to use the skill for very broad concepts like 'work-productivity,' 'security,' and 'analysis,' which are common across many unrelated requests. This can cause unintended activation and routing of user tasks into the wrong skill, creating prompt-scope confusion and increasing the chance that sensitive or safety-critical requests are handled by an overbroad workflow.
The keyword list includes extremely generic terms such as 'security,' 'gate,' 'every,' 'must,' 'pass,' and 'before,' which are likely to appear in many benign conversations. If used as activation signals, they can spuriously invoke this skill in unrelated contexts, leading to incorrect task handling, policy bypass by misrouting, or interference with more appropriate specialized skills.
The skill description says it should activate for very broad topics like work-productivity, security, and generic needs for workflows, checklists, analysis, or implementation support. This can cause the skill to trigger in many unrelated conversations, increasing the chance of unintended instruction injection, user confusion, or the wrong workflow being applied to sensitive requests.
The keyword list includes generic words such as "security," "must," "pass," "before," and "every," which are common in normal conversation. Such unconstrained triggers make accidental activation highly likely and expand the skill's reach into unrelated or sensitive contexts where its guidance may be inappropriate.
The trigger sentence is broad enough to match ordinary requests about 'security', 'gate', 'every', or 'activate', which can cause the skill to activate outside its intended niche. In an agent ecosystem, unintended activation can override more appropriate skills, route sensitive tasks through the wrong workflow, and create confusion or unsafe handling of user intent.
The trigger phrases are broad, generic, and include common terms like 'help me', 'I need', 'security', and 'bug fix', which can cause the skill to activate for unrelated everyday requests. In an agent ecosystem, overbroad activation can route sensitive or irrelevant tasks into this skill unexpectedly, creating prompt-scope confusion and increasing the chance of unsafe or unintended workflow execution.
The trigger keywords and phrases are excessively broad, including generic terms like "security," "every," "must," "pass," "before," and "bug fix." In an agent ecosystem, this can cause unintended activation on ordinary user requests, leading the skill to intercept unrelated tasks or security-sensitive workflows and potentially alter agent behavior in ways the user did not intend.
The example triggers model activation on vague, marketing-style language rather than precise operational intent, reinforcing an overly permissive trigger scope. This makes accidental invocation more likely and weakens downstream assurances that the skill was selected because it truly matches the user's need.
The skill content and instructions are presented entirely in Simplified Chinese, and the filename indicates a zh-CN locale, but there is no natural-language statement offering the user a language choice or opt-in. This can violate language/locale policy when the skill implicitly constrains interaction to a single language without documented justification.
The example trigger phrases use broad openings like "Help me" and "I need a practical workflow," which do not define clear trigger boundaries. This reinforces permissive matching behavior and may train integrators or routers to invoke the skill on vague, underspecified requests.
The default_prompt is overly broad and uses generic everyday phrasing such as 'help me' and wide topical language, which increases the chance that unrelated user requests will accidentally match and invoke this skill. In a security-oriented workflow helper, unintended invocation can inject unneeded security guidance or workflow behavior into unrelated conversations, creating confusion, overreach, or prompt-surface abuse opportunities.
Enabling allow_implicit_invocation without clear trigger constraints allows the platform to auto-select this skill based on ambiguous language rather than explicit user intent. Because this skill is framed around security workflows and practical implementation help, accidental invocation can cause unintended actions, irrelevant security recommendations, or expanded attack surface through opportunistic routing.
This trigger reads like generic user intent rather than a clearly bounded invocation condition, so many unrelated productivity or bug-fix requests could match it. That ambiguity increases the chance of accidental routing, lowering reliability and potentially causing users to receive security-oriented workflow behavior when they asked for something else.
The file lists specific language instruction and guide files as English and Chinese variants, but does not state that users may choose their preferred language or that language selection is optional. This can be read as imposing fixed locales rather than explicitly offering a user-directed language choice.
This file is presented as a Chinese README (README.zh-CN.md), but key user-facing descriptive content is written in English without indicating whether the user can choose language output. That creates a locale-policy concern because the skill presentation implicitly forces mixed-language content rather than clearly honoring the user's language context.
