Security audit
Security checks for vulnerabilities and agentic risk
The skill is a documentation-only helper with broad activation wording, but it does not contain hidden code, credential handling, persistence, or destructive behavior.
This appears safe to install as a lightweight workflow helper. The main practical issue is that its triggers are too broad, so users or maintainers may want to invoke it explicitly or narrow/disable implicit invocation to avoid it being selected for ordinary software or data questions.
The trigger phrases are so broad and generic that the skill may activate for ordinary software/data requests that were not intended for this specific capability. In an agent ecosystem, this can cause inappropriate routing, accidental invocation, and increased exposure to downstream unsafe or irrelevant behavior, especially because the skill positions itself as applicable to many adjacent workflows.
The skill description and invocation guidance are broad enough that the skill may activate for loosely related requests instead of clearly scoped AdMapix/raw-data developer tasks. Overbroad routing can cause incorrect skill selection, resulting in irrelevant guidance, accidental access to sensitive context, or bypass of more appropriate specialized safeguards.
The keyword list contains generic terms such as 'raw', 'data', 'layer', 'apps', and 'revenue' that are common across many unrelated prompts. Such generic triggers significantly increase the chance of unintended invocation, which can misroute user requests and make the agent apply this skill in contexts where its assumptions, instructions, or safety posture do not fit.
The trigger sentences are so generic that they can match ordinary user requests that merely mention needing help or a practical workflow, causing unintended activation of this skill. In an agent ecosystem, overly broad invocation criteria can route unrelated tasks into the wrong skill, leading to mis-execution, confusing outputs, or unsafe automation if downstream actions are taken based on an incorrect match.
The trigger sentences and keywords are broad, repetitive, and partially generic, which increases the chance the skill is invoked for loosely related requests rather than clear AdMapix-specific workflows. Unintended invocation can cause the agent to apply the wrong workflow or expose users to instructions and artifacts outside their intent, reducing safety and reliability.
The file descriptions explicitly label this document and related files as Chinese and English variants, and this README itself is presented in Chinese without any visible language choice or opt-in guidance for users. Under the policy, forcing a specific language without user choice can be a locale-policy violation unless the constraint is clearly justified.
The example trigger sentences demonstrate activation from very broad natural-language phrasing without clarifying limits, reinforcing loose matching behavior. This makes accidental or excessive invocation more likely and trains maintainers toward permissive trigger design that can interfere with correct skill routing and contextual safety controls.
The activation condition in the description is vague and expansive, stating the skill should be used whenever a user asks for broad categories like software-and-data, admapix, raw, or data layer support. Ambiguous enablement criteria can lead to over-triggering and unintended selection of this skill, which is a security and reliability issue because it broadens the attack surface for prompt-driven behavior.
The trigger keywords include very broad terms such as "raw", "data", "layer", and "apps", which can easily appear in unrelated user requests. This can cause accidental invocation of the skill in contexts where it is not intended, increasing the chance of irrelevant behavior, prompt-surface expansion, or routing users into a workflow with mismatched assumptions.
The default prompt invokes the skill with a very broad natural-language phrase covering generic terms like software, data, workflow, analysis, and implementation support. Because implicit invocation is enabled, this overlap with ordinary user requests can cause the skill to activate unintentionally, expanding its influence over unrelated conversations and increasing the chance of prompt-routing mistakes or unsafe task handling.
The trigger scope is ambiguous because it mixes broad keywords ('software-and-data', 'raw', 'data', 'apps') with partially truncated natural-language trigger sentences, without clear activation boundaries. This increases the chance of false-positive selection and skill confusion, especially in multi-skill environments where routing decisions may be automated and rely heavily on keyword overlap.
The example trigger phrases use generic help-seeking language like "Help me" and "I need a practical workflow," without contrasting negative examples. This can train or bias routing systems toward matching ordinary assistance requests that are unrelated to the skill, causing accidental activation and confusion.
