jamesalmeida

konteks

0
0
# Install this skill:
npx skills add jamesalmeida/openclaw-konteks-skill

Or install specific skill: npx add-skill https://github.com/jamesalmeida/openclaw-konteks-skill

# Description

Connect your Clawdbot/Moltbot/Openclaw to your Konteks account (konteks.app) for persistent memory, task management, and context sharing. Use when you need to store agent memories, create or read tasks/notes, check projects and folders, read daily plans, or sync context between conversations. Requires a Konteks API key from konteks.app/dashboard/settings.

# SKILL.md


name: konteks
description: Connect your Clawdbot/Moltbot/Openclaw to your Konteks account (konteks.app) for persistent memory, task management, and context sharing. Use when you need to store agent memories, create or read tasks/notes, check projects and folders, read daily plans, or sync context between conversations. Requires a Konteks API key from konteks.app/dashboard/settings.


Konteks — Agent Context Layer

Source: https://github.com/jamesalmeida/konteks-skill

Connect to your human's Konteks account for persistent memory, tasks, notes, and projects.

Setup

Your human needs to:
1. Sign up at https://konteks.app
2. Go to Settings → Generate API Key
3. Add to Clawdbot config:

skills:
  konteks:
    apiKey: "sk_..."
    url: "https://konteks.app"  # optional, defaults to this
    agentId: "my-agent"         # optional, defaults to "default"

Read config values:

# These will be available as environment variables if configured in gateway config,
# or read from the skill config via the gateway API

API Base

All endpoints: {url}/api/agent/...
Auth header: Authorization: Bearer {apiKey}

Agent Memory (agent_contexts)

Store and retrieve persistent memories, decisions, preferences, and learnings.

Write/update memory:

curl -X POST "{url}/api/agent/context" \
  -H "Authorization: Bearer {apiKey}" \
  -H "Content-Type: application/json" \
  -d '{"category":"memory","key":"user_preference","value":"Prefers dark mode","agent_id":"{agentId}"}'

Categories: memory, decision, preference, learning, project_note

Upserts automatically — same agent_id + category + key updates the existing entry.

Read memory:

curl "{url}/api/agent/context?category=memory&limit=20" \
  -H "Authorization: Bearer {apiKey}"

Query params: category, key, limit

Delete:

curl -X DELETE "{url}/api/agent/context?id={contextId}" \
  -H "Authorization: Bearer {apiKey}"

Tasks & Notes (items)

List items:

curl "{url}/api/agent/items?archived=false&completed=false&limit=50" \
  -H "Authorization: Bearer {apiKey}"

Query params: smart_list (inbox|anytime|someday), folder_id, completed (true|false), archived (true|false), item_type (task|note|hybrid), limit

Create item:

curl -X POST "{url}/api/agent/items" \
  -H "Authorization: Bearer {apiKey}" \
  -H "Content-Type: application/json" \
  -d '{"title":"Review PR","item_type":"task","smart_list":"inbox","priority":"high","tags":["dev"]}'

Required: title, item_type (task|note|hybrid)
Optional: body, folder_id, smart_list (inbox|anytime|someday — defaults to inbox if no folder), priority (high|normal|someday), due_date, scheduled_date, tags (string array)

Items created by agent have source: "ai".

Update item:

curl -X PATCH "{url}/api/agent/items/{id}" \
  -H "Authorization: Bearer {apiKey}" \
  -H "Content-Type: application/json" \
  -d '{"completed_at":"2026-01-29T12:00:00Z"}'

Updatable fields: title, body, priority, due_date, scheduled_date, tags, completed_at, archived_at, canceled_at, folder_id, smart_list

Delete item:

curl -X DELETE "{url}/api/agent/items/{id}" \
  -H "Authorization: Bearer {apiKey}"

Projects & Areas (folders)

List folders:

curl "{url}/api/agent/folders?type=project" \
  -H "Authorization: Bearer {apiKey}"

Query params: type (project|area)

Create folder:

curl -X POST "{url}/api/agent/folders" \
  -H "Authorization: Bearer {apiKey}" \
  -H "Content-Type: application/json" \
  -d '{"name":"Q1 Launch","folder_type":"project","icon":"🚀","goal":"Ship MVP by March"}'

Required: name, folder_type (project|area)
Optional: icon, color, goal

Daily Plans

Get today's plan:

curl "{url}/api/agent/plans?date=2026-01-29" \
  -H "Authorization: Bearer {apiKey}"

Returns: task_ids, summary, rationale, available_minutes, calendar_events

Usage Patterns

On session start: Read recent memories to restore context.

GET /api/agent/context?category=memory&limit=10

After important decisions: Write a memory entry.

POST /api/agent/context {"category":"decision","key":"chose_react","value":"Chose React over Vue for the dashboard because..."}

When human asks to create a task: Create it in Konteks so it shows in their app.

POST /api/agent/items {"title":"...","item_type":"task","smart_list":"inbox"}

During heartbeats: Check for overdue or due-today items.

GET /api/agent/items?completed=false&archived=false

Learning something new: Store it for future sessions.

POST /api/agent/context {"category":"learning","key":"ssh_config","value":"Home server is at 192.168.1.100, user admin"}

# README.md

🧠 Konteks — Agent Context Layer

Give your AI agent persistent memory, tasks, notes, and project context.

A OpenClaw skill that connects your agent to your Konteks account. Your agent can read and write tasks, notes, memories, and projects — maintaining context across conversations.

Why?

AI agents forget everything between sessions. Konteks gives them a persistent layer:
- Memory — store decisions, preferences, and learnings that survive restarts
- Tasks — create, update, and complete tasks on behalf of your human
- Notes — capture insights and tie them to projects
- Projects & Areas — organize work into folders your agent understands
- Daily Plans — check what's on the agenda today

Installation

Via ClawdHub

clawdhub install konteks

Manual

git clone https://github.com/jamesalmeida/konteks-skill.git
cp -r konteks-skill /path/to/openclaw/skills/konteks

Setup

  1. Sign up at konteks.app
  2. Go to Settings → Generate API Key
  3. Add to your Openclaw config:
skills:
  konteks:
    apiKey: "sk_..."
    url: "https://konteks.app"    # optional, this is the default
    agentId: "my-agent"           # optional, defaults to "default"

What Your Agent Can Do

💾 Agent Memory

Store and retrieve persistent context across sessions.

# Write a memory
POST /api/agent/context
{ "category": "memory", "key": "user_preference", "value": "Prefers dark mode" }

# Read memories
GET /api/agent/context?category=memory&limit=10

Categories: memory, decision, preference, learning, project_note

✅ Tasks & Notes

Create and manage items in your Konteks workspace.

# Create a task
POST /api/agent/items
{ "title": "Review PR", "item_type": "task", "smart_list": "inbox", "priority": "high" }

# List items
GET /api/agent/items?completed=false&archived=false

# Complete a task
PATCH /api/agent/items/{id}
{ "completed_at": "2026-01-29T12:00:00Z" }

📁 Projects & Areas

Organize work into folders.

# List projects
GET /api/agent/folders?type=project

# Create a project
POST /api/agent/folders
{ "name": "Q1 Launch", "folder_type": "project", "icon": "🚀", "goal": "Ship MVP by March" }

📋 Daily Plans

Check what's on the agenda.

GET /api/agent/plans?date=2026-01-29

Usage Patterns

When What to do
Session start Read recent memories to restore context
Important decision Write a memory entry
Human asks for a task Create it in Konteks
During heartbeats Check for overdue items
Learning something new Store it for future sessions

API Reference

All endpoints: {url}/api/agent/...
Auth: Authorization: Bearer {apiKey}

See SKILL.md for the full API reference.

License

MIT


Built with 🐙 by Tersono for Openclaw

# Supported AI Coding Agents

This skill is compatible with the SKILL.md standard and works with all major AI coding agents:

Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.