candidate-kb
Manage candidate knowledge base. Use when ingesting CVs, adding profile information, or querying candidate details for resume/cover letter generation.
An MCP server that provides on-demand skill loading for AI coding assistants.
---
name: kvcore-mcp-cli
description: Use this skill when users need KVcore CRM actions through MCP/CLI (contacts, tags, notes, calls, email, text, campaigns), including raw endpoint access and optional Twilio call fallback.
---
# KVcore MCP/CLI Skill
Use this skill for KVcore CRM operations from chat interfaces.
## Environment
Required:
- `KVCORE_API_TOKEN`
Optional:
- `KVCORE_BASE_URL` (default `https://api.kvcore.com`)
- `KVCORE_TIMEOUT_MS`
- `TWILIO_ACCOUNT_SID`, `TWILIO_AUTH_TOKEN`, `TWILIO_FROM_NUMBER`
## MCP Server
Start with:
```bash
npm run dev:kvcore-mcp
```
Primary tools:
- `kvcore_contact_search`, `kvcore_contact_get`, `kvcore_contact_create`, `kvcore_contact_update`
- `kvcore_contact_tag_add`, `kvcore_contact_tag_remove`
- `kvcore_note_add`
- `kvcore_call_log`, `kvcore_call_schedule`
- `kvcore_email_send`, `kvcore_text_send`
- `kvcore_user_tasks`, `kvcore_user_calls`
- `kvcore_campaigns_refresh`
- `kvcore_request` (raw endpoint access)
- `twilio_call_create` (fallback outbound call)
## CLI
Build:
```bash
npm run build
```
Examples:
```bash
node packages/kvcore-cli/dist/index.js contact search --query "john smith" --pretty
node packages/kvcore-cli/dist/index.js email:send --contact-id 123 --subject "Quick update" --body "Following up" --pretty
node packages/kvcore-cli/dist/index.js text:send --contact-id 123 --body "Can we connect today?" --pretty
node packages/kvcore-cli/dist/index.js call:schedule --json '{"contact_id":123,"user_id":456,"scheduled_at":"2026-02-15 10:00:00"}' --pretty
node packages/kvcore-cli/dist/index.js call:twilio --to "+14165550001" --twiml "<Response><Say>Hello</Say></Response>" --pretty
```
## Scope Notes
KVcore Public API v2 supports contacts, notes, call logging, send email/text, schedule call, user task/call listing, and campaign refresh.
For endpoints not wrapped yet, use `kvcore_request` or CLI `raw`.
---
name: kvcore-mcp-cli
description: Use this skill when users need KVcore CRM actions through MCP/CLI (contacts, tags, notes, calls, email, text, campaigns), including raw endpoint access and optional Twilio call fallback.
---
# KVcore MCP/CLI Skill
Use this skill for KVcore CRM operations from chat interfaces.
## Environment
Required:
- `KVCORE_API_TOKEN`
Optional:
- `KVCORE_BASE_URL` (default `https://api.kvcore.com`)
- `KVCORE_TIMEOUT_MS`
- `TWILIO_ACCOUNT_SID`, `TWILIO_AUTH_TOKEN`, `TWILIO_FROM_NUMBER`
## MCP Server
Start with:
```bash
npm run dev:kvcore-mcp
```
Primary tools:
- `kvcore_contact_search`, `kvcore_contact_get`, `kvcore_contact_create`, `kvcore_contact_update`
- `kvcore_contact_tag_add`, `kvcore_contact_tag_remove`
- `kvcore_note_add`
- `kvcore_call_log`, `kvcore_call_schedule`
- `kvcore_email_send`, `kvcore_text_send`
- `kvcore_user_tasks`, `kvcore_user_calls`
- `kvcore_campaigns_refresh`
- `kvcore_request` (raw endpoint access)
- `twilio_call_create` (fallback outbound call)
## CLI
Build:
```bash
npm run build
```
Examples:
```bash
node packages/kvcore-cli/dist/index.js contact search --query "john smith" --pretty
node packages/kvcore-cli/dist/index.js email:send --contact-id 123 --subject "Quick update" --body "Following up" --pretty
node packages/kvcore-cli/dist/index.js text:send --contact-id 123 --body "Can we connect today?" --pretty
node packages/kvcore-cli/dist/index.js call:schedule --json '{"contact_id":123,"user_id":456,"scheduled_at":"2026-02-15 10:00:00"}' --pretty
node packages/kvcore-cli/dist/index.js call:twilio --to "+14165550001" --twiml "<Response><Say>Hello</Say></Response>" --pretty
```
## Scope Notes
KVcore Public API v2 supports contacts, notes, call logging, send email/text, schedule call, user task/call listing, and campaign refresh.
For endpoints not wrapped yet, use `kvcore_request` or CLI `raw`.
Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.