intent-guardian_skill

This skill helps you maintain focus by tracking your task stack and gently nudging you back to the current task after interruptions.
  • Python

2.5k

GitHub Stars

4

Bundled Files

2 months ago

Catalog Refreshed

3 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

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

  • _meta.json284 B
  • clawhub.json1.0 KB
  • README.md2.7 KB
  • SKILL.md12.3 KB

Overview

This skill watches your desktop activity and maintains a real-time task stack to preserve your train of thought. It detects when interruptions cause you to forget an in-progress task and sends gentle, context-aware reminders so you can resume work quickly. The agent also learns long-term focus patterns to personalize detection and reminder timing.

How this skill works

It continuously senses active window titles, application names, and optional richer signals (ActivityWatch or periodic screen summaries) to build an activity stream. An LLM segments that stream into tasks, tracks a task stack (active, suspended, completed), and applies heuristics plus a learned focus profile to detect likely forgetting. When forgetting is detected, it issues non-intrusive reminders that can include the reason you left and helpful context to resume the original task.

When to use it

  • You frequently context-switch between apps and lose track of your prior task
  • You finish a quick reply or lookup and then wonder “what was I doing?”
  • You want an always-on focus companion that nudges you back without interrupting work
  • You want to learn long-term patterns about interruptions and peak focus windows
  • You want reminders that carry forward the result of why you left (so you can finish faster)

Best practices

  • Enable ActivityWatch for richer, more accurate window and URL context
  • Configure excluded apps and working hours to respect privacy and prevent over-notification
  • Tune interruption_threshold_minutes and reminder_cooldown_minutes to match your attention span
  • Allow screen capture only with a trusted vision model and confirm screenshots are transient
  • Review and edit local JSON logs periodically to correct inferred intents and improve personalization

Example use cases

  • After a Slack ping, get a subtle prompt reminding you you were mid-way through a product spec and offering the lookup result
  • Daily focus report summarizing interruptions, forgotten tasks, and peak focus hours for planning
  • Auto-suppress reminders during declared deep-work sessions or when Focus mode is on
  • Convert suspended task items into calendar or reminder entries if you choose to delegate them
  • Use the task stack to answer questions like “What was I doing before I got sidetracked?”

FAQ

Yes. Activity logs, task stacks, and focus profiles are stored locally as human-readable JSON and controlled by your configuration.

Will it nag me with too many reminders?

No. The skill respects reminder cooldowns, working hours, and learns which reminders you accept versus dismiss to reduce unnecessary nudges.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational