Intent Guardian for Openclaw

An intelligent focus companion that monitors desktop activity to maintain a real-time task stack and remind you of forgotten work after interruptions.

huilingp
v1.0.0
Feb 26, 2026
0
1.1k
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install intent-guardian

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install intent-guardian using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Intent Guardian?

Intent Guardian is a specialized focus assistant designed to mitigate the cognitive cost of context switching. By continuously sensing active window titles and application usage, it builds a dynamic model of your current work intent. As a key component in the ecosystem of Openclaw Skills, it helps knowledge workers recover the approximately 23 minutes typically lost after an interruption by maintaining a persistent task stack.

The skill goes beyond simple activity logging by using LLMs to infer the purpose of your actions. It detects when you have been derailed—for example, by a Slack message or a quick browser search— and haven't returned to your primary objective. By storing state in a structured JSON stack, it can offer context-aware nudges that facilitate a seamless return to deep work.

Intent Guardian Use Cases

  • Automatically tracking work context across multiple applications like VSCode, Chrome, and Slack.
  • Receiving gentle reminders to return to a high-priority document after being sidetracked by notifications.
  • Monitoring personal focus patterns to identify which applications cause the highest rate of task abandonment.
  • Maintaining a history of intent for daily stand-ups or focus reporting.

How Intent Guardian Works

  1. The sensing layer executes scripts to poll the active window name and application at a configurable interval.
  2. Raw activity data is appended to a local activity log for processing.
  3. During the agent heartbeat, an LLM analyzes the recent log to segment activities into logical tasks and infer the user's current intent.
  4. The task stack is updated, moving previous tasks to a suspended state if a new context is detected.
  5. The agent monitors the duration of suspended tasks against a forgetting threshold based on your focus profile.
  6. If forgetting is detected, the agent proactively sends a reminder via the system notification channel or integrated chat.
  7. User interactions with reminders are logged to refine future detection sensitivity and focus metrics.

Intent Guardian Setup

Install the skill via the command line:

npx playbooks add skill openclaw/skills --skill intent-guardian

To enable continuous monitoring, add the following logic to your HEARTBEAT.md file:

## Intent Guardian Check
1. Read activity from memory/skills/intent-guardian/activity_log.jsonl
2. Update task_stack.json with inferred intents
3. Trigger reminders if suspended tasks exceed threshold

For enhanced accuracy, you can optionally install ActivityWatch and set activitywatch_enabled: true in the configuration.

Intent Guardian Data Schema & Taxonomy

The skill organizes its intelligence across several local files to ensure privacy and persistence:

File Purpose
task_stack.json A structured list of active and suspended tasks with start times and completion estimates.
activity_log.jsonl A raw stream of window titles and application names captured by the sensors.
focus_profile.json A learned profile containing peak focus hours, average duration, and interruption source analytics.

Intent Guardian Advanced Features

  • ActivityWatch integration for rich, document-level tracking and URL monitoring.
  • Vision-based sensing using models like Claude or Gemini to analyze screen content for deeper context.
  • Custom interruption heuristics that distinguish between deliberate task abandonment and accidental forgetting.
  • Personalized focus profiling that adapts Openclaw Skills behavior based on your unique work rhythm.
  • Automated daily reporting via cron integration to summarize focus effectiveness and interruption sources.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0