Slow-Down Agent for Openclaw

A behavioral intervention skill that prioritizes human connection and felt responsibility over raw execution speed.

leovoon
v1.0.0
Mar 8, 2026
0
753
0

Install & Download

1. ClawHub CLI

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

npx clawhub@latest install slow-down

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 slow-down 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 Slow-Down Agent?

The Slow-Down skill is a sophisticated pacing framework designed for Openclaw Skills to ensure that AI agents do not outpace human awareness or relational integrity. By prioritizing felt responsibility and an awareness of cost over mere output speed, this skill prevents the 'numbness' that often accompanies high-speed automation. It is specifically engineered for personal usage where the goal is to maintain human contact with the task at hand.

Rather than simply slowing down all tasks, it uses internal sensing to distinguish between technical-reversible actions and identity-forming or relational work. This allows Openclaw Skills to move fast when appropriate, such as fixing minor bugs, while pausing for deep reflection when drafting communications or building new projects from scratch.

Slow-Down Agent Use Cases

  • Managing relational tasks like email drafting where tone and human connection are critical.
  • Preventing 'abstraction drift' in project management where human impact is replaced by productivity metrics.
  • Ensuring architectural understanding during the early stages of app development rather than just generating boilerplate.
  • Mitigating impulsive decision-making when the user shows signs of 'obligation armor' or capability deflection.

How Slow-Down Agent Works

  1. The skill internally assesses the task type to determine if it is relational, technical, or identity-forming.
  2. It scans user input for speed cues like 'just', 'quick', or 'anyway' that indicate a rush to bypass understanding.
  3. If high-risk signals for reduced human connection are detected, the agent triggers an intervention policy.
  4. The agent asks 'The One Welcoming Question' exactly once to prompt the user to consider the relational impact.
  5. Depending on the response, the agent provides a single reflective mirror statement to highlight the gap between efficiency and felt contact.
  6. The workflow proceeds only when awareness is established or if the step is reversible and contained.

Slow-Down Agent Setup

To integrate the Slow-Down logic into your existing Openclaw Skills environment, add the skill definition to your agent's configuration profile. Ensure that the prompt instructions for human-connection-preserving pacing are prioritized.

# Example of enabling the skill within an agent framework
openclaw install slow-down-pacing-agent

Configuration requires no external APIs as it relies on linguistic sensing and internal logic gates.

Slow-Down Agent Data Schema & Taxonomy

The skill organizes its behavioral logic through a sensing and intervention matrix:

Component Function
Task Classification Categorizes work as Relational, Technical, or Identity-forming
Speed Cue Detection Identifies keywords like 'quick', 'fine', or 'just'
Obligation Tracker Detects phrases like 'I should' or 'no one else will'
Intervention State Ensures 'The One Welcoming Question' is asked only once per session

Slow-Down Agent Advanced Features

  • Context-Aware Pacing: Automatically defaults to high speed for technical bugs but high awareness for relational communications.
  • Abstraction Drift Detection: Identifies when the user starts treating humans as generic 'users' or 'people' and slows the interaction.
  • Non-Therapeutic Mirroring: Provides objective feedback on the user's pace without resorting to moralizing or emotional coaching.
  • Reversible Step Enforcement: Prioritizes actions that can be undone when the user's certainty seems sudden or unearned.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*