Project Explorer

Files

Story files

Browse the notes, research, decisions, and workshops used to develop the series.

Browse the files used to develop the story.

Search 577 documents about the world, characters, plot, research, decisions, workshops, and earlier ideas.

Current document skills/README.md

Shared AI Skills

This folder contains reusable methods for any AI agent working in the Seeds of the Throne vault. Skills may guide writing style, research, story development, continuity review, public-site updates, or other repeatable work.

Skills are tool-neutral. An agent may use its own tools, but it should preserve the workflow, constraints, and output standards defined by the skill.

How agents discover skills

  1. Review the catalog below.
  2. Match the current request to a skill description.
  3. Read that skill's complete SKILL.md before acting.
  4. Read only the referenced materials needed for the current task.
  5. Follow the author's request and root AGENTS.md if either conflicts with a skill.

Do not load every skill automatically. Progressive loading keeps the working context focused.

Folder standard

skills/
  skill-name/
    SKILL.md
    references/    optional detailed guidance
    scripts/       optional repeatable automation
    assets/        optional templates or output resources

Each skill name uses lowercase hyphenated words. Each SKILL.md begins with YAML frontmatter containing only name and description.

Skill catalog

  • 0: explore story ideas, preserve alternatives, record decisions, and update memory deliberately.
  • 0: investigate real-world mechanisms with sourced uncertainty and no automatic canon.
  • 0: audit chronology, character knowledge, causality, systems, terminology, and status.
  • 0: coach fiction craft by default and provide short, non-canon demonstrations when the author explicitly asks for samples.
  • 0: modular story-writing system for scenes and chapters, including character voice, dialogue, exposition, suspense, controlled human variance, development and line editing, development-residue cleanup, pressure-turn diagnosis, preference routing, anti-AI checks, research/question triage, customizable style controls, a 100-point evaluation rubric, and regression benchmarks while preserving author and canon authority.
  • 0: publish approved developments to the static story atlas and verify the result.
  • 0: create, evaluate, and store consistent character art and story imagery through a model-neutral visual registry and approved reference library.
  • 0: design, prototype, implement, and visually review the story atlas and Project Explorer through an author-gated coded-prototype workflow.

The initial story-workflow set was structurally validated and forward-tested on 2026-08-11. The image system was added and validated on 2026-08-12. The coaching skill was added on 2026-08-15. The prose skill was expanded into a modular writing and regression-testing system on 2026-08-18 and refined with controlled human variance on 2026-08-19. The coded website-design skill was added on 2026-09-08 after the author rejected a prose-and-mockup-led visual implementation. Test results are recorded in 07 QA/Shared Skill Tests.md.

Creating a skill

Copy skill-template.md into skills/<skill-name>/SKILL.md, replace every placeholder, and keep the instructions concise. Put long examples or detailed domain material in references/ and link them directly from SKILL.md.

Test a new skill on a realistic task before treating it as dependable.