Konrad stands under Samuel's hidden red control while Sylvan observes from the clear opposing side.

Project Explorer

Seeds of the Throne

Inspect the corrected story, trace unresolved causes, and work through source-linked brainstorming packets. The repository browser remains available below.

Open the workshop workspace Browse the project

Development workspace · September 6 workshop pass

Find the cause. Test the choice.

All twenty workshop gates have accepted directions. Their unresolved mechanics remain visible. The active Story Completion gate remains SC-010 Question 7.

Public, read-only development material with full spoilers. Draft answers stay in your browser until exported; they do not update the vault. A spoiler boundary is not privacy protection.

01 · Foundation

The environment creates the bond.

The leaders developed Luminai within an interactive colonization environment. Sylvan tests a more deeply integrated generation after thousands of years of development.

Inspect the controlling premise
02 · Institutional gap

Control must protect people.

Final-years control is established. Stopping authority, cumulative harm, privacy, and remedies still need concrete rules.

Test the leaders' safeguards
03 · Causal result

Autonomy becomes access.

Konrad's Daemon verifies an apparent separation. Konrad authorizes reactivation, which attaches his revived system to Samuel's hidden hierarchy. The false evidence still needs definition.

Inspect the takeover result
04 · Two-planet endgame

George goes where Samuel cannot.

Samuel and the older criminals remain on the previous containment planet. George participates directly on Sylvan's world, then becomes Samuel's intended scapegoat.

Inspect George's role
Read Astra's integrated assessmentOpen Current PickupOpen the reader's atlas

Story completion · Live from the vault

From architecture to manuscript.

The story advances horizontally: every active problem receives the same level of development before any one branch moves deeper.

Integrated foundation review · September 5, 2026

The environment develops the bond. The outcome tests responsibility.

The leaders created an interactive colonization environment and developed Luminai within it. Konrad tries to prove his experienced daemon superior; Samuel turns his reactivation into access. Sylvan already has decisive control during the final years. Exact safeguards and presentation mechanics remain open.

Current method Bound action, observe the method, compare the record, expose the hidden command.

Read the foundation update record

9 / 27story tasks complete

Current checklist completion33%
33%
Current sweep
Macro Shape
Current task
SC-010
  1. 01 Macro
  2. 02 Causal
  3. 03 Agency
  4. 04 Systems + evidence
  5. 05 Sequence
  6. 06 Scene map
  7. 07 Scene development
  8. 08 Draft

Current working fronts

  • Human–Luminai breakthrough
  • Evidence chain
  • Character agency
Open the full story roadmap

Project explorer

Browse the repository.

Search or browse 539 documents across canon, development systems, story loops, prose tools, public work, and session history.

3.3 KB Updated View source on GitHub

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.

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. 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.