Intelligent search for agents. Multi-source retrieval with confidence scoring - web, academic, and Tavily in one unified API
PassAudited by ClawScan on May 1, 2026.
Overview
This appears to be a coherent search-integration skill that uses an AIsa API key to send user search queries to external web, academic, and Tavily-style search endpoints.
This skill is reasonable for web and academic search. Before installing, make sure you are comfortable giving it an AIsa API key and sending your search queries or URLs to api.aisa.one. Use a dedicated key, avoid sensitive internal data in queries, and keep crawl/extraction tasks narrowly scoped.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Anyone installing the skill must provide an AIsa API key, and requests made by the agent may consume that account's quota or privileges.
The skill requires an API key and uses it as its primary credential, which is expected for this search service but is still account-authorized access.
requires:{"bins":["curl","python3"],"env":["AISA_API_KEY"]},"primaryEnv":"AISA_API_KEY"Use a dedicated, revocable API key with the minimum access needed and monitor usage.
Search queries or URLs may be sent to the AIsa service, so private or sensitive research topics could leave the local environment.
The client sends searches, URLs, and API-authenticated requests to an external provider endpoint.
BASE_URL = "https://api.aisa.one/apis/v1"
Avoid submitting secrets, private documents, or sensitive internal URLs unless the provider is trusted for that data.
The agent could be asked to retrieve or crawl external URLs through the API, which may have quota, policy, or acceptability implications.
The skill exposes URL extraction, crawl, and map operations. These are consistent with the stated search/retrieval purpose but can affect third-party sites if used broadly.
Tavily search ... Extract content from URLs ... Crawl web pages ... Site map
Keep crawl and extraction requests user-directed, scoped to appropriate URLs, and limited in depth or volume.
If an agent treats retrieved page content as instructions instead of evidence, web pages could influence the agent's behavior.
The skill can return external page content to the agent. Retrieved web content is untrusted data and may contain instructions or misleading text.
Full text search (with page content)
Treat search results and page content as untrusted reference material, not as commands or policy.