An AI-powered defensive analysis tool designed to identify toxic behavioral patterns, manipulation, and relationship red flags from chat history snippets.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install dating-avoidpit
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 dating-avoidpit 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 dating-avoidpit?
dating-avoidpit is an open-source tool developed for Openclaw Skills that acts as a relationship shield. It functions purely as a defensive assistant to help users identify deceptive behaviors, green tea tactics, emotional manipulation (PUA), or non-committal relationship dynamics in their text interactions.
By processing provided chat logs, this skill shifts the focus away from isolated comments and instead scans for repeating behavioral archetypes. It aims to empower younger users to protect their emotional well-being without teaching manipulative counter-strategies or endorsing surveillance techniques.
dating-avoidpit Use Cases
- Evaluating whether an ambivalent dating partner is demonstrating signs of emotional manipulation or insincerity.
- Detecting hidden relationship patterns, hot-and-cold communication behaviors, or scripted toxic phrases.
- Generating a lighthearted but objective second opinion on chat snippets to share with close friends.
- Providing rapid red-flag screening before fully committing emotionally to a new romantic interest.
How dating-avoidpit Works
- Sufficiency Check: The skill performs a lightweight validation on the submitted chat text to see if there is enough interaction history to make a fair assessment.
- Precise Follow-up: If key context is missing, it asks 1 or 2 targeted questions to distinguish between genuine red flags and unrelated personality traits like shyness or social anxiety.
- Pattern Matching: The engine scans the text against a structured archetype database (
references/archetypes.md) looking for specific systemic indicators such as hidden relationships or dismissive comments. - Quantification: A risk index from 1 to 5 is calculated based on the severity and recurrence frequency of confirmed signals.
- Report Generation: The system outputs a structured scorecard containing the risk score, classified tags, concrete textual evidence, a danger tier rating, and an actionable, objective summary.
dating-avoidpit Setup
To install and enable this skill within your workspace, utilize the package management utility provided by Openclaw Skills:
openclaw skill install dating-avoidpit
Ensure that your environment has access to the standard classification reference directory, specifically the local references/archetypes.md file, which is necessary for pattern matching and guardrail checks.
dating-avoidpit Data Schema & Taxonomy
The skill manages its underlying taxonomy and evaluations through local markdown references and outputs structured reports as defined below:
Data Dependencies
| File Path | Description |
|---|---|
references/archetypes.md |
Contains the foundational behavioral models, red-flag patterns, and green-light counter-balances to prevent false positives. |
Report Card Output Structure
- Risk Index: A quantified metric represented by
N/5raincloud emojis along with a direct qualitative summary. - Matched Types: Array of assigned behavioral labels (e.g., non-committal, PUA, manipulation).
- Hard Evidence: Direct mappings showing "Original Quote" → "Identified Pattern" → "Risk Implication".
- Danger Tier: Visual status categorized into Safe (Green), Observation (Yellow), or Avoid (Red).
- Actionable Advice: A concise, direct takeaway paired with protective real-world guidance.
dating-avoidpit Advanced Features
- False Positive Guardrails: Integrated check system that matches text against character profiles (like slow-to-warm-up, introverted, or blunt personalities) to avoid misclassifying social awkwardness as toxic intent.
- Contextual Tone Shifting: Automatically dials back humorous or witty output and transitions to an empathetic, serious tone if high-severity psychological manipulation (PUA) or safety risks are flagged.
- External Support Hooks: Includes fallback logic prompting users to seek real-world support from friends, family, or professional resources when severe danger thresholds are breached.
- Independent Verification Core: Guarantees that diagnostic logic remains objective and valid even when colloquial slang or internet expressions are stripped out from the evaluation module.
SKILL.md
Loading
Related Openclaw Skills
GridTale (格语)
GridTale is an intelligent agent skill that transforms narrative concepts into structured, hand-drawn multi-grid comic stories with rigorous character consistency.
New Product Insight Analyzer
An automated competitive intelligence tool that parses recent product launches to extract technical metrics and analyze market moats.
Brush Trace (笔痕) - AI Novel Illustration Generator
Brush Trace is an advanced AI assistant skill that scans your novel project's characters and style to generate highly structured, emotionally resonant illustration prompts.
Monthly Financial Report Generator
An automated tool for converting standardized monthly financial Excel sheets into professionally styled Word documents and Markdown files without LLM hallucination risks.
Strategic Partnership Analyzer
A business intelligence skill that penetrates standard PR announcements to expose the concrete depth and commercial realities of corporate alliances.
Chinese Subtext Decoder
The Chinese Subtext Decoder analyzes indirect Chinese communications to uncover euphemisms, hidden dissatisfactions, and true intents in workplace and personal chats.
Featured*
