knowledge-bases

Manage Frontline knowledge bases, sources, crawls, and semantic search via CLI.

Updated Jul 27, 2026
One-click install
npx skills add https://github.com/api-evangelist/frontline --skill knowledge-bases-api-evangelist
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: knowledge-bases
Source: https://github.com/api-evangelist/frontline/tree/main/skills/vendor/knowledge-bases
Command: npx skills add https://github.com/api-evangelist/frontline --skill knowledge-bases-api-evangelist

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Managing knowledge bases through a web UI is slow when you need to script or automate the lifecycle of indexed content. This Skill lets you create, update, and delete Frontline knowledge bases, add URL and website-crawl sources, track sync history, run semantic search, and assign KBs to agents entirely from the command line. ## Core Features & Use Cases - Knowledge base CRUD: Create, list, update, and delete shared or private KBs with paginated JSON or table output. - Source management: Add URL sources, start website crawls with depth and link limits, resync sources, and inspect crawl sync history with status filters. - Semantic search and agent assignment: Query indexed content with similarity scores and assign shared KBs to agents by ID. - Use Case: You want an agent to answer questions from your public docs. Create a KB, start a crawl of your docs site with a max-links limit, monitor the sync history until it completes, then assign the KB to the agent. ## Quick Start Ask the AI to create a shared knowledge base named Help Center, crawl https://docs.example.com with up to 50 links, and assign it to your agent.

Frequently Asked Questions about knowledge-bases

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a knowledge base from the command line?▼

Run frontline kb create with a --name and optional --description flag, adding --private to make it non-shared. The command returns the new KB as JSON, and you can list all KBs with frontline kb list --table.

How do I add website content to a knowledge base?▼

Use frontline kb source add-crawl with the KB id, a root --url, and a required --max-links value between 1 and 1000. Optionally set --max-depth, --include-paths, and --exclude-paths to control which pages get indexed.

Can I upload files to a Frontline knowledge base via the CLI?▼

No, file uploads are not available through the CLI. Sources are limited to URLs and website crawls, which are fetched and indexed asynchronously after being added.

Why does knowledge base search return empty results?▼

A KB with no indexed sources yet returns an empty results array. Check that your sources finished processing by inspecting frontline kb source list or the sync history before running frontline kb search.

How do I resync a website crawl source?▼

Run frontline kb source resync with the KB id and --website-id set to the website source id. Use --source-ids instead when resyncing individual URL or web-page sources; exactly one of the two flags is required.

Why can't I assign a knowledge base to an agent?▼

Only shared KBs can be assigned to agents; private KBs are rejected. Also note that frontline kb assign replaces the agent's full set of assigned KBs, so include every KB id you want to keep.