Defines clear ownership boundaries for decisions to establish whether the user provides input, the AI proposes options, or the AI acts autonomously.
Horismos is a specialized epistemic protocol for Claude Code designed to eliminate ambiguity in human-AI collaboration by explicitly defining decision-making responsibility. By generating a structured BoundaryMap, it classifies project domains into specific ownership categories—User-Supplied, AI-Proposes, or AI-Autonomous—ensuring that Claude neither oversteps its authority nor stalls due to lack of direction. This skill is essential for maintaining project integrity and user control during complex development tasks where the lines of responsibility might otherwise be blurred.
主な機能
01Classifies domains into User-Supplied, AI-Proposes, or AI-Autonomous categories
0288 GitHub stars
03Generates a comprehensive BoundaryMap to track and enforce epistemic ownership
04Scans codebases and configurations for boundary signals to enrich decision context
05Ensures 'Definition over Assumption' by resolving ownership before execution proceeds
06Proactively identifies undefined decision domains within any given task scope
ユースケース
01Resolving delegation uncertainty during multi-step refactoring or system migrations
02Defining collaboration levels when onboarding Claude to a large or unfamiliar codebase
03Setting clear decision boundaries at the start of complex architectural implementations