Cleanup: retire the dead name Yukejulhel from canon

Data-quality cleanup, not new lore. The author confirms Yukejulhel is an obsolete pre-rename name for Medesjul (species.firgol.sentient.julhel.medesjul, canon_json), left over from WorldAnvil and carried into the JSON import. Four canon files still list it as a related species: countries_and_continents.continents.silgolt.nations.izheim.izheim_union, ...kokit.duchy_of_kokit, ...nasun.nasun_ground, ...hihn.hihn_confluence (all canon_json). No entity exists at any Yukejulhel id, so every one of these edges is currently unresolved.

Canonized 9/9 items resolved

Proposal workspace — not canon. Accepting an item records your review decision. Writing it into canon requires a separate exact preview and typed confirmation. Overview items are never imported.

Review the whole package · 1 items

Connected-lore projection

Relationships

Switch between family, organization, diplomacy, causal, and general lenses.

No family relationships in this package.

Selected item

Package overview

Accepted Overviewv1

Overview frames this package for review. It is never written to canon.

Manual revision

Title Text

# Ask

This package fixes a data bug rather than adding lore. **Yukejulhel is a dead name for Medesjul** — an earlier naming that was superseded before the JSON import and never swept out of the WorldAnvil-derived content.

Four canon country files still list `Yukejulhel` in `related_species`:

- `...silgolt.nations.izheim.izheim_union`
- `...silgolt.nations.kokit.duchy_of_kokit`
- `...silgolt.nations.nasun.nasun_ground`
- `...silgolt.nations.hihn.hihn_confluence`

Because no entity exists at any Yukejulhel id, all four of those edges currently dangle — they point at nothing. Meanwhile Medesjul, the actual majority population of most of the western cluster, is absent from the related_species of every one of those countries. The relation graph is wrong in both directions at once.

## How this surfaced

Badly, and I want to be straight about it. While writing the western-cluster packages I flagged `Kaetijul` vs `Kaetijulhel` as a canon inconsistency; you corrected me — `-jul` is just the short form of `-julhel`. I then reported "Yukejulhel has no species entity" as a genuine remaining gap and wrote it into eight package overviews. That was also wrong, for a different reason: it has no entity because it *is* Medesjul.

Two false alarms in a row on the same axis. The lesson I'd draw is that I was pattern-matching on suffixes instead of asking what the names actually referred to, and I should have asked after the first correction rather than generating a second theory.

The entity payloads in the ten Silgolt packages have all been corrected to use **Medesjul**, so nothing in the queue will write the dead name into a new canon file on promotion.

## Items

Eight `relationship` edits — for each of the four affected countries, remove the dangling `Yukejulhel` edge and add the correct `Medesjul` one. These are surgical and independently reviewable; accept or reject them per country.

## Open questions for the human reviewer

- **Prose occurrences are not covered here.** Kokit, Nasun, and Hihn also use "Yukejulhel" inside their `content` body text. Patching those means replacing whole content sections, which I did not want to do blind on live canon files — a wrong keystroke there costs more than the bug. Say the word and I'll read each file and file precise `field_patch` items for the prose.
- **Long form or short?** Those four files use the long forms throughout (`Kaetijulhel`). The literal long form of Medesjul would be *Medesjulhel*, which appears nowhere in canon. I have used **Medesjul**, matching the species entity's own name. If you'd rather the country files stay internally long-form, these edges should say Medesjulhel instead — your call, and worth settling before anything is promoted.
- Worth a grep of the wider corpus. If the dead name reached these four files, it may also be sitting in `writing/` or in WorldAnvil exports not yet imported.

What changed?

AI action · one call

Action Personality Provider Model Instruction

Overview accepted as read. Nothing to canonize.