Horror Night for Openclaw

An immersive AI-driven horror mystery game where players survive a snow mountain lodge through investigation and strategic choices.

ouwenjie03
v1.0.0
Jan 27, 2026
1
2.6k
0

Install & Download

1. ClawHub CLI

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

npx clawhub@latest install banshee-s-last-cry

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 banshee-s-last-cry 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 Horror Night?

Horror Night is a sophisticated interactive fiction skill designed for AI agents, offering a deep psychological thriller experience. Set in an isolated snow mountain lodge during a blizzard, players are thrust into a serial murder mystery with six strangers. This skill utilizes Openclaw Skills to manage a complex narrative environment where every decision impacts the survival of the group.

The game stands out by moving away from rigid multiple-choice menus in favor of natural language exploration. By integrating Openclaw Skills, the AI agent can process organic player actions, maintain consistent character relationships, and manage a dynamic clue system that dictates the flow of seven distinct chapters.

Horror Night Use Cases

  • Engaging in a solo interactive horror adventure with an AI narrator.
  • Hosting AI-powered script-kill (Jubensha) sessions for digital roleplay.
  • Demonstrating advanced state management in narrative-driven Openclaw Skills.
  • Training AI models in suspenseful storytelling and logical consistency.

How Horror Night Works

  1. The AI initializes the environment by loading core reference files for game state, character networks, and clue hierarchies.
  2. The player starts the prologue, which establishes the setting and the first incident at the lodge.
  3. The agent monitors player input in natural language, comparing actions against the hidden clue system and character trust levels.
  4. Narrative progression occurs in segments of 200-400 words, ensuring a focused and atmospheric experience.
  5. The agent updates a persistent JSON state file to track survival, discovered evidence, and branch triggers.
  6. The story concludes in one of seven unique endings based on the cumulative data tracked by the Openclaw Skills framework.

Horror Night Setup

To begin using this skill, ensure your AI agent has access to the references directory and execute the following sequence:

# Load core system logic
read references/system_game_state.md
read references/system_clue_system.md
read references/system_character_info.md

# Initialize the first chapter
read references/ch00_start.md

The AI will then take over as the narrator, managing state through the Openclaw Skills protocol.

Horror Night Data Schema & Taxonomy

The skill organizes its narrative and logic through a structured metadata system to ensure consistency across sessions:

Component Format Description
Game State JSON Tracks current chapter, scene ID, and unlocked narrative branches.
Clue System Markdown/List Categorizes evidence into Surface, Hidden, and Core layers.
NPC Profiles Markdown Defines 6 unique characters with trust levels and secret motives.
Endings Logic Gate Triggers one of 7 final states based on survival and truth metrics.

Horror Night Advanced Features

  • Progressive storytelling that pauses for player interaction to prevent information overload.
  • Dynamic NPC relationship mapping where dialogue choices affect trust and story outcomes.
  • Layered clue mechanics requiring active investigation rather than passive reading.
  • Automatic state persistence via Openclaw Skills to allow for multi-session play.
  • Non-linear chapter branching including specific paths for solo exploration or team cooperation.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*