Migliorate llm_short_description per tutti i connector — tool names espliciti
Ogni connector.json ora ha una llm_short_description che elenca i tool specifici che espone, così l'LLM sa esattamente cosa attivare (lazy load). SKALD.md: documentata criticità del campo llm_short_description nel system prompt dell'LLM.
This commit is contained in:
@@ -16,7 +16,7 @@
|
||||
{
|
||||
"lang": "en",
|
||||
"description": "AI-powered web search, content extraction, and embeddings via Exa's neural search engine. Perfect for deep research, finding specific information, and analyzing page content. Works without an API key (free rate-limited tier) \u2014 add your own key for higher limits.",
|
||||
"llm_short_description": "Search the web and extract page content using Exa's neural search engine. Tools: web_search_exa, web_fetch_exa."
|
||||
"llm_short_description": "Exa neural search engine: semantic web search and content extraction. Tools: web_search_exa (semantic/keyword search with domain filter, date range, result count), web_fetch_exa (extract full content + metadata from URLs). Best for semantic search that understands intent, not just keywords. Free tier available without API key."
|
||||
}
|
||||
],
|
||||
"auth": {
|
||||
|
||||
Reference in New Issue
Block a user