Security audit
Security checks for vulnerabilities and agentic risk
This is a documentation-help skill for OpenAPI/Swagger work, with no executable code or hidden system access, though its auto-invocation wording is broader than ideal.
Installers should be aware that the skill may activate for loosely related API or developer-experience requests because its triggers are broad. Review or tighten the trigger text if precise routing matters, but there is no evidence of hidden execution, data exfiltration, credential use, or persistence.
The trigger sentence is overly broad and can activate on ordinary user phrasing rather than a clear, scoped request for this specific skill. That creates misrouting and prompt-hijack surface area by causing the agent to invoke the skill in contexts where OpenAPI documentation help was not actually intended.
The trigger sentences are broad, malformed, and close to generic user phrasing, which increases the chance the skill is invoked unintentionally in unrelated conversations. In an agent environment, unintended invocation can route user input into the wrong workflow, causing confusing outputs, context leakage across tasks, or execution of documentation-generation behavior when not requested.
The trigger phrases are broad and generic enough to match routine requests about APIs, documentation, or workflow help, which can cause the skill to activate outside its intended scope. Over-broad activation increases the chance of unintended prompt injection exposure, user confusion, or inappropriate routing to this skill when another tool or safer default behavior should apply.
The skill description and activation guidance are broad enough that the skill may trigger for generic software or API-related requests outside its intended scope. Over-broad routing can cause mis-selection of the skill, leading to irrelevant guidance, instruction shadowing, or increased exposure to untrusted artifact-generation paths when another narrower skill should have handled the request.
The keyword triggers include broad terms such as 'openapi', 'swagger', and especially 'developer experience', which can appear in many harmless or unrelated contexts. This increases the chance of accidental activation and prompt-surface expansion, where the skill injects its workflow into conversations that do not actually require OpenAPI documentation support.
The trigger description is broad and includes common terms like OpenAPI, Swagger, API documentation, and REST API, which can appear in many normal engineering conversations. This increases the chance the skill is invoked when the user did not explicitly want it, causing scope confusion, incorrect routing, or unintended use of the skill's workflow.
The example trigger phrases use generic expressions such as 'Help me' and 'I need a practical workflow' combined with a long requirement description, but they do not establish clear invocation boundaries. In systems that rely on example-driven routing, such broad phrasing can train or encourage over-triggering on ordinary requests, leading to accidental skill activation and reduced reliability.
The skill enables implicit invocation, but its description and default prompt are broad and not narrowly scoped to precise trigger conditions. This can cause the agent to invoke the skill in loosely related contexts, increasing the chance of unintended data exposure, prompt-scope expansion, or misuse of the skill without clear user intent.
The activation examples are ambiguous and do not clearly bound when the skill should or should not run. Ambiguous routing rules can cause unintended invocation, which may override a better-matched skill, expose irrelevant instructions, or let adversarial phrasing steer execution into this skill.
The file mixes English user guidance with multiple Chinese-language evidence titles and links, but does not explain whether the skill is intended to operate in a specific language or locale. This can create an implicit language/locale bias without user opt-in or a documented regional justification.
The file descriptions explicitly label separate English and Chinese documents, and this file itself is a Chinese-localized README, but there is no natural-language statement offering language choice or explaining locale selection. Under the policy, forcing a specific language without user opt-in can be a locale-policy violation unless the constraint is clearly documented and justified.
The example trigger sentences are malformed and truncated, so they do not provide reliable invocation guidance for either users or routing systems. Poor examples can cause inconsistent activation behavior and make it harder to distinguish legitimate use from accidental or adversarial prompt phrasing.
No suspicious patterns detected.
