Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-only product validation planning skill with broad activation wording but no hidden code, persistence, credential access, or external execution.
Install this only if you want broad product-validation and startup-planning help. Expect it may activate on general startup, SaaS, prototype, or validation requests because its triggers are broad; review outputs before using them for business decisions because the included demand evidence appears noisy and should not be treated as independent market proof.
The trigger phrases are extremely broad and can match common requests about workflows, practical help, or product ideas, which increases the chance that the skill activates outside its intended scope. Over-broad invocation can route unrelated user requests into business-planning behavior, causing prompt/skill hijacking of normal conversations and reducing trust in skill selection boundaries.
The trigger phrases are broad help-seeking prompts such as 'Help me' and 'I need a practical workflow', combined with common product terms. This can cause the skill to activate in situations far beyond the intended scope, effectively hijacking unrelated user requests and increasing the chance that the agent routes sensitive business or planning conversations into this skill without clear user intent.
The skill description is broad enough to match many ordinary business, planning, workflow, and implementation requests, which can cause the agent to invoke this skill outside its narrowly intended use. Over-broad activation increases the attack surface for prompt hijacking, tool misuse, or unintended routing because users may be funneled into a high-authority workflow when they only asked for general help.
The trigger keywords are highly generic terms like 'validation,' 'prototype,' 'saas,' and 'startup,' which commonly appear in normal conversation and unrelated tasks. This makes accidental or adversarial triggering easier, allowing the skill to capture requests it was not designed to handle and potentially override better-scoped system behavior.
The skill description and activation guidance are broad enough to match many ordinary business, product, and startup conversations, which can cause the skill to activate outside its intended scope. Over-broad routing is dangerous because it may override more appropriate domain-specific skills, steer users into unintended workflows, or increase the chance that adversarial prompts trigger this skill opportunistically.
The keyword list contains highly generic terms like 'validation', 'prototype', 'saas', and 'startup', which are common across many unrelated requests. This creates a prompt-routing vulnerability where routine conversations can unintentionally invoke the skill, causing misclassification, noisy agent behavior, and increased exposure to prompt-injection attempts through unnecessary skill loading.
The default prompt uses a very broad invocation phrase tied to a wide set of user intents, which can cause the skill to activate in situations beyond narrowly intended product-validation tasks. Because implicit invocation is enabled, this increases the chance of over-triggering, context capture, and unintended routing of user requests into this skill without clear user intent.
The trigger sentence is broad enough to match many ordinary product or business-assistance requests, which can cause the skill to activate outside its intended scope. Overbroad activation increases the chance of prompt routing mistakes, unintended disclosure of irrelevant instructions, and interference with more appropriate skills, especially in multi-skill agent environments.
This trigger lacks clear activation boundaries because it accepts a generic request for a 'practical workflow' without enough contextual qualifiers. In an agent system, ambiguous triggers can be exploited or accidentally invoked by unrelated requests, causing inappropriate tool selection and degraded reliability of downstream outputs.
The file lists separate English and Chinese instruction and guide files, but does not state how the user's preferred language is selected or that language choice is optional. This can create an implicit language/locale behavior without explicit user opt-in.
This file is a Chinese-language README ("README.zh-CN.md") and does not state that language selection is optional or user-driven. Under the policy, locale-specific behavior should either be explicitly opt-in or clearly justified as region-specific.
