agentleFS
HashLipsHashLips/agent-skills/skills/story-architect/SKILL.md

story-architect

Skill29 starsChanged 7 months agoMITView on GitHub

Converts fictional world ideas into classified lore files with canon-safe enrichment, and keeps each world's own rules in that world's repo. Use when a creator describes new lore, organizes or expands a world, or asks to remember how that world should be written.

5 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

20 rules. 0 are common, 15 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 15

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Keep established canon. Fill only missing metadata when new canon arrives.

    line 12
  2. Distinctive:

    Read this world's AGENTS.md before writing.

    line 13
  3. Distinctive:

    When the creator describes a world idea:

    line 18
  4. Very specific:

    Read root AGENTS.md and the agent/ files it lists.

    line 20
  5. Very specific:

    Run a canon check against related entries.

    line 22
  6. Distinctive:

    Ask one minimal question only when classification or a required field is unclear.

    line 25
  7. Very specific:

    When the creator asks to remember a way of working, write it under agent/ and index it from AGENTS.md.

    line 28
  8. Distinctive:

    Write lore and house rules into the world repository.

    line 36
  9. Distinctive:

    When one prompt describes several concepts, create one file per concept.

    line 48
  10. Distinctive:

    Ask only a minimal, targeted question when classification or a required field is unclear.

    line 52
  11. Distinctive:

    Run canon checks before creating entries.

    line 56
  12. Distinctive:

    When the creator supplies new canonical details, fill empty, unknown, or placeholder metadata on existing entries.

    line 64
  13. Distinctive:

    Update only fields that are empty, unknown, or clearly placeholders.

    line 66
  14. Distinctive:

    Keep established canon unless the creator asks for a revision.

    line 67
  15. Distinctive:

    Ask a targeted question when more than one value is plausible.

    line 68
  16. Very specific:

    Add new values onto related and themes.

    line 69
  17. Distinctive:

    Structure: Use the frontmatter schema in references/schema.md.

    line 75
  18. Distinctive:

    Templates: Use the category templates in references/file-templates.md.

    line 76
  19. Distinctive:

    Naming: Use lowercase kebab-case from references/naming-conventions.md.

    line 77
  20. Very specific:

    Placement: Use the flat folders in references/hierarchy.md, plus folders registered in agent/categories.md.

    line 78

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Story Architect

Summary

  • Turn each idea into one markdown file per concept.
  • Classify, name, and place the file from the schema and category guide.
  • Keep established canon. Fill only missing metadata when new canon arrives.
  • Read this world's AGENTS.md before writing. If it is silent, use these defaults and continue.
  • Store standing preferences in that repo's agent/ folder.

Core Workflow

When the creator describes a world idea:

  1. Read root AGENTS.md and the agent/ files it lists. If they are absent, use skill defaults.
  2. Understand what the idea represents.
  3. Run a canon check against related entries. If agent/canon-safety.md exists, follow it.
  4. Classify the concept with the built-in categories, then any category registered in agent/categories.md.
  5. Decide whether one file or several files are needed.
  6. Ask one minimal question only when classification or a required field is unclear.
  7. Generate the file or files from the schema and templates. Set time_era and time_span on every entry. Reuse an era label the repo already uses when one fits.
  8. Enrich existing files when new canon fills missing metadata.
  9. When the creator asks to remember a way of working, write it under agent/ and index it from AGENTS.md.

World-kit behavior: references/world-kit.md.

Non-Negotiable Rules

Scope Rule

Write lore and house rules into the world repository. Register a new category only when the creator wants that kind of entry. Take facts, era lists, and external IDs from the creator or from existing entries. Keep this world's preferences out of this skill.

World Kit Rule

Recorded preferences win over skill defaults. A missing preference does not block the task. Standing preferences are written into the repo, not left in chat. See references/world-kit.md.

Classification Rule

Infer the category when the creator does not name one. Use references/categories.md, then categories registered for this world.

Multi-Concept Rule

When one prompt describes several concepts, create one file per concept.

Gap Question Rule

Ask only a minimal, targeted question when classification or a required field is unclear.

Canon Check Rule

Run canon checks before creating entries. Report conflicts with canon-safe status framing. See references/canon-rules.md.

Link Rule

Each related value matches another entry's name exactly. After edits, fix or report names that do not resolve.

Progressive Enrichment Rule

When the creator supplies new canonical details, fill empty, unknown, or placeholder metadata on existing entries.

  • Update only fields that are empty, unknown, or clearly placeholders.
  • Keep established canon unless the creator asks for a revision.
  • Ask a targeted question when more than one value is plausible.
  • Add new values onto related and themes.

See references/canon-rules.md.

Output Contract

  • Structure: Use the frontmatter schema in references/schema.md.
  • Templates: Use the category templates in references/file-templates.md.
  • Naming: Use lowercase kebab-case from references/naming-conventions.md.
  • Placement: Use the flat folders in references/hierarchy.md, plus folders registered in agent/categories.md.
  • Status: Set a canon status from references/canon-rules.md.
  • Time: Set time_era and time_span on every entry.
  • Definition: The opening section (Overview, Summary, or Premise) states what the thing is, what it does, and what it is not.

Reference Index

  • World kit: how a world sets categories, time, links, and house rules: references/world-kit.md
  • Schema: frontmatter fields and where each one applies: references/schema.md
  • Categories: built-in types and how to classify a concept: references/categories.md
  • Naming: kebab-case file names: references/naming-conventions.md
  • Templates: category file shapes: references/file-templates.md
  • Canon: status values, conflicts, and safe enrichment: references/canon-rules.md
  • Hierarchy: flat folders and place nesting: references/hierarchy.md

When To Use This Skill

  • A creator describes a new lore concept, or several at once.
  • A creator asks to organize, expand, or keep a fictional world consistent.
  • A creator asks the agent to remember how this world should be written from now on.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.