Oh-story-claudecode is a skill pack for web fiction writing that covers the complete workflow of long-form and short-form novel creation, from market analysis through final delivery.
The tool addresses the challenge of writing commercially viable web fiction by codifying professional author methodology into three core steps: analyzing trending works to understand genre patterns and character archetypes, deconstructing successful narratives to extract reusable plot modules and pacing structures, and applying proven techniques like hooks and tension management. The approach treats narrative patterns as deterministic emotional satisfaction, organizing the writing process around four dimensions: reverse-engineering bestsellers, modularizing plot components for recombination, managing narrative state across context layers, and human-AI collaboration.
The skill pack suits writers working within the web fiction ecosystem who want systematic guidance on commercial appeal and structural consistency. It distinguishes between long-form and short-form workflows, with the short-form variant emphasizing scene-level decision points rather than mechanical quotas for dialogue proportion or hook spacing. The tool integrates with multiple AI platforms including Claude Code, Google Antigravity, OpenCode, ZCode, OpenClaw, Codex CLI, and Reasonix, and can be deployed to any web-based AI environment that reads project files. Installation happens through direct instruction to supported platforms or command-line setup.
The project maintains active refinement of its writing methodology. Recent updates removed mechanical constraints from short-form scene construction in favor of outcome-based evaluation, introduced cross-session author memory to maintain consistency across conversations, added reference material gatekeeping to prevent contamination during composition phases, and unified stylistic priority handling while fixing reference consumption issues. The tool requires redeployment and new conversation sessions after updates to reflect methodology changes, with version tracking maintained through an agents_version parameter.
