Security audit
Security checks for vulnerabilities and agentic risk
This SEO research skill is mostly coherent and transparent about using an AIsa API key, crawling public websites, and sending SEO inputs to AIsa/DataForSEO, though its invocation wording is broader than ideal.
Install only if you are comfortable providing an AIsa API key and sending SEO research inputs, crawled public page text, competitor domains, and prompts to AIsa/DataForSEO. Use it for SEO keyword research tasks, not general browsing or private content extraction, and avoid crawling internal, authenticated, customer, or confidential pages.
The skill clearly instructs the agent to use environment variables, read and write local files, and make network requests, yet it does not declare explicit permissions for those capabilities. This creates a trust and review gap: a harness or user may invoke the skill without understanding that local content and site data will be transmitted to external services and written to disk.
The invocation guidance goes beyond narrowly scoped SEO keyword research and explicitly says to use the skill when the user needs generic web search, research, source discovery, or content extraction. That broad trigger can cause the agent to select this skill for unrelated tasks, increasing the chance of unnecessary crawling, external API use, and data handling outside the intended security and policy boundaries.
The trigger condition is overly broad because it includes generic 'web search, research, source discovery, or content extraction' tasks, which can cause the skill to activate outside narrow SEO keyword research. In context, that matters because activation leads to crawling websites and sending gathered content to external AIsa/DataForSEO endpoints, increasing the chance of unnecessary data collection and exfiltration.
The skill directs the agent to transmit crawled site data and keyword research inputs to external AIsa/DataForSEO services. In this skill's context that behavior is expected, but it is still security-relevant because crawled pages may contain sensitive business information, internal strategy, customer data, or accidentally included secrets that would leave the local environment.
Use these AIsa endpoints:
- Data APIs: `https://api.aisa.one/apis/v1/...`
- LLM gateway: `https://api.aisa.one/v1/chat/completions`
Never print or commit API keys. If the key is missing, ask the user to set `AISA_API_KEY`.
The LLM gateway is another external transmission path, distinct from the data API, and may receive prompts containing extracted website content, competitor lists, or internal strategy notes. Even if the provider is trusted, the skill lacks clear limits on what portions of crawled or local data can be forwarded, creating avoidable exposure risk.
Use these AIsa endpoints:
- Data APIs: `https://api.aisa.one/apis/v1/...`
- LLM gateway: `https://api.aisa.one/v1/chat/completions`
Never print or commit API keys. If the key is missing, ask the user to set `AISA_API_KEY`.
No suspicious patterns detected.