An intelligent emotional monitoring skill that detects AI-directed frustration and appends grounding reminders to help users reset and maintain productivity.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install calm-down
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install calm-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 Anxiety Detector?
The Anxiety Detector is a sophisticated emotional-logic layer designed for AI coding agents and assistants. It monitors real-time conversational signals to identify when a user is experiencing high levels of frustration or cognitive fatigue specifically directed at the AI's performance. By recognizing these emotional collapse patterns, the skill ensures the agent remains a helpful collaborator rather than a source of escalating stress.
Unlike generic wellness bots, this Openclaw Skills extension prioritizes technical utility first. It follows a strict Solve-Acknowledge-Ground protocol, ensuring the user's immediate technical problem is addressed before offering a brief, human-centric suggestion to take a mental break. This approach maintains professional boundaries while significantly improving the long-term developer experience.
Anxiety Detector Use Cases
- High-pressure coding sessions where repetitive AI errors lead to user venting or insults.
- Late-night development marathons (00:00–05:00) where cognitive load is high and frustration triggers are frequent.
- Situations involving repeated complaints about persistent issues or memory lapses in the LLM.
- Identifying sentence structure collapse or excessive punctuation as indicators of user burnout.
How Anxiety Detector Works
- Signal Monitoring: The skill scans incoming messages for two or more frustration markers, such as insults, all-caps text, or expressions of hopelessness.
- Context Validation: It distinguishes between frustration directed at the AI versus external stressors or polite factual corrections to avoid false positives.
- Problem Resolution: The agent first generates a technical solution to the user's query to prove continued competence.
- Failure Acknowledgment: A single, concise sentence is generated to own any AI error without making excuses.
- Grounding Append: A randomized, non-robotic reminder from the grounding bank is appended to the very end of the response, suggesting a short physical break.
Anxiety Detector Setup
To integrate this skill into your environment, ensure you have the Openclaw Skills framework initialized. Install the detector using the following CLI command:
openclaw install anxiety-detector
Configure your config.yaml to include the grounding language preferences and ensure the system has access to the user's local time for temporal trigger detection.
Anxiety Detector Data Schema & Taxonomy
The Anxiety Detector organizes its logic and signal data through a structured internal taxonomy to ensure precise triggers.
| Data Component | Description | Examples |
|---|---|---|
| Trigger Logic | Boolean gates requiring 2+ active signals | is_nighttime, has_insults, is_venting |
| Grounding Bank | A localized collection of string templates | Short_Punchy, Flowing_Reflective |
| Response Map | The 3-part sequence for output generation | [Solve] -> [Acknowledge] -> [Ground] |
| Language Detection | Matches response language to the user's dominant input | en, es, fr, etc. |
Anxiety Detector Advanced Features
- Temporal Contextualization: Automatically increases sensitivity during late-night hours (00:00–05:00) to account for sleep deprivation.
- Multi-Lingual Grounding: Automatically detects the user's language and selects a culturally and linguistically appropriate reminder from the bank.
- Non-Repetition Algorithm: Ensures that the same grounding reminder is never repeated twice in a row for a more authentic, human feel.
- Signal Reference Auditing: Uses
signal-examples.mdas a baseline for continuous improvement of trigger accuracy within Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
OpenClaw User Onboarding
A guided onboarding experience that transforms a fresh OpenClaw installation into a personalized, feature-aware AI assistant.
Memory Curator
A specialized tool for reviewing, pruning, and syncing structured AI memory to maintain context quality and efficiency.
gog
A comprehensive Google Workspace CLI tool that enables AI agents to interact with Gmail, Calendar, Drive, Sheets, and Docs through a unified interface.
Huo15 Office Doc
An enterprise-grade Word document generation skill that automates professional formatting, versioning, and template matching.
Explore Mode Agent
A specialized read-only agent for systematic codebase investigation and architectural research.
Agent YML Generator
A high-precision tool that transforms business requirements and SOPs into Dify 1.3.1 compatible Agent workflow YAML files.
Featured*
