Rokugan RPG fantasy setting conversion for D&DNext (preliminary thoughts)

[Li Shenron]

Rokugan is my favourite oriental setting, and it has its own RPG rules system developed by AEG, therefore it won't ever have an official D&DNext version, but 5e is the system I'm looking forward to DM and play with in the following years. Therefore I just know that if I want to play a 5e game in Rokugan, it's inevitable I have to write my own conversion.

I won't really start until D&DNext is actually released and the core books are in my hands, but I have been gathering some preliminary thoughts about what to do. I don't know at this moment if I'll be allowed to share the conversion to the public.

In many ways, Rokugan and D&D don't go together well, at least because the latter is traditionally focused on exploration and combat while Rokugan has a lot more going on in the social environment. While that is also often part of a lot of D&D games, there are characters in Rokugan which are almost entirely designed around that. Furthermore, D&D combat is based on HP attrition, while Rokugan combat is more lethal, rather similar to save-or-die. Finally, the spellcasting system in Rokugan is significantly different from vancian magic.

Nevertheless, my own experience with Rokugan is entirely confined to its 3e/d20 version, so I'll be fine with recreating that unorthodox version of it.

You are welcome to comments on my thoughts and ideas!

 

[Li Shenron]

Preliminary thoughts: essential setting elements

Here I'm rambling a bit about what are elements of the Rokugan fantasy setting that IMO need to be reflected in the converted rules, and generate some "design marks".

When I'll start making the actual conversion, I'll restrict myself to a "basic Rokugan campaign". What does that mean?

It means essentially to focus on the most iconic characters only, which is what typically you should start from in your first game in Rokugan. In the world of Rokugan, the norm for a player character is to hail from the Rokugani noble caste (let's call them "bushi" rather than "samurai" to avoid confusion with the Samurai class), and belong to either a Great Clan, a Minor Clan, an Imperial Family, or be a Ronin. Non-noble human characters (peasants or slaves), non-human characters and "gaijin" (foreigners) are not basic characters and interact very differently with everybody else.

Design mark #1: only the Human race in the basic campaign (Rokugani bushi)
Design mark #2: focus on Great Clans first, leave Minor Clans, Imperial Family and Ronin for a later stage

Furthermore, all bushi train with a School since young age, and normally stick with it forever; there is a stern division between martial schools (Samurai), magic schools (Shugenja), court schools (Courtiers) and monasteries (Monks). Almost everything that can be learned, has to be learned from a master, but masters don't "sell" training (money doesn't work for Rokugani bushi). There are options to learn specific things from another School or master, but there is no such thing as freely shifting from school to school or choosing to learn anything you want (in fact, it could almost be said that there is no individual freedom at all in Rokugan). This causes stern multiclassing restrictions between the main classes, however in d20 other classes work more like "service classes" that can be freely taken to complement the character mechanics, without the need to have a narrative meaning (example: you are a Samurai, you're trained in one specific martial School, mechanically you might be a Samurai/Fighter/Ranger/Barbarian if everything is appropriate to your School, narratively you're still just a Samurai of that School).

Design mark #3: only Samurai, Shugenja, Courtier and Monk classes needed in the basic campaign
Design mark #4: no multiclassing

Honor is essential for a Rokugan campaign, while Alignment is unimportant and mostly left as a roleplay aid. To make Honor a more tangible element of the game, an actual Honor rules system should be there. Fortunately, the d20 Honor system is pretty much plug-and-play and can probably be used in 5e with little or no adjustment.

Design mark #5: just reuse the d20 Honor system

Skills (anything non-martial and non-magical) are very important in Rokugan. The original d20 versions of the 4 main classes all had at least 4 skill points per level, so all human PCs could max out 5 skills minimum. D&DNext characters use broader skills than in 3e, and all have at least 4 of them, so maybe not much change is needed here. However the skills list need to be filtered for potentially dishonorable skills, and more Rokugani-specific skills need to be added. Backgrounds in 5e represent your "job" or "social role" when not adventuring, but all Rokugan basic PC are nobles, and don't have a "job"; most sample Backgrounds are inappropriate, and the few left would make the PC too rigid, thus it's best to go with the "pick your own skills" option.

Design mark #6: remove dishonorable (sub)skills from the list
Design mark #7: identify what iconic Rokugani skills are missing from the list
Design mark #8: no backgrounds, freely pick your trait+skills+tools

Clan, Family and Ancestor are Rokugan-specific choices essential in character design. Clan in d20 didn't grant benefits straightforward, but instead it determined what you can learn (feats, skills, spells, prestige classes etc.). The choice of Family typically determined your favored class (not applicable), an additional class skill or small skill bonus, and extra equipment. Your chosen Ancestor was your first level bonus feat, called an "Ancestor feat", which is something that each PC has one and only one (you had to pick one Ancestor feat at 1st level, you could never have another, and there is no other way to gain Ancestor feats); this means that despite being "feats" they are completely separate from all other feats in the game, hence there is no need to convert them into 5e feats, they can remain their own separate thing.

Design mark #8: there has to be plenty of clan-restricted character material (feats, spells, subclasses, equipment)
Design mark #9: Ancestor feats remain a separate subsystem

Magic in Rokugan works differently than in typical D&D. The Shugenja is the only PC class who does magic, which is a little bit of everything: magic is related to the spirit-based religion of Rokugan, so narratively the Shugenja is partly a Mage, partly a Cleric and partly a Druid. That's not a problem, anyway it's the only spellcasting class. But the spellcasting system used by the Legend of the Five Rings RPG is very different from vancian. The d20 version of the Shugenja used D&D spells and slots, but at least cast spells as a Sorcerer. What is most appropriate for 5e still needs to be thought out carefully.

Another thing that needs to be in, is Void Use, which is about supernatural deeds but potentially open to every character. In d20, most abilities related to Void Use were therefore feats. This may or may not be the best solution in D&DNext, I still have to think about it.

As a last remark, as long as a "basic Rokugan campaign" is concerned, there is no need to worry about less-than-straightforward characters, let's say "characters with issues" such as Ronin, Ninja, and tainted characters. The DM is definitely going to need something to represent Maho-Tsukai or Maho-Bujin and other adversaries, but these could be designed ad-hoc, without generalized rules for the time being.

...on second thoughts, probably Taint should be there already in the basic game. The d20 Taint system is probably good as-is except for the numbers involved, which might need some adaptation to fit bounded accuracy.

Design mark #10: reuse the d20 Taint system, with adjusted numbers

 

[Li Shenron]

Preliminary thoughts: Races

In a basic Rokugan campaign, your only possible choice of race is Human.

This causes one concern: what sense does it make if every PC has +1 to all abilities?

At the same time, the game needs for humans (and humans only, or more precisely non-foreign humans from the noble caste) to belong to one Clan, and to one Family of such clan. Also, each human needs to have an Ancestor, although this does not necessarily depend on Clan and Family.

So instead of adding these three choices, I will replace the +1 to all abilities with the choice of Clan, Family and Ancestor.

Clan in d20 does not grant immediate benefits, but typically determines which Schools you can choose to be trained at, which in turns "unlock" the option of picking certain feats, spells, special class abilities, prestige classes etc. This can pretty much work the same way in D&DNext.

Family in d20 determined the following:

- your favoured class: this doesn't work in 5e and is removed
- your starting honor range: this is based on the Honor system which can be ported as-is from d20
- one extra class skill: this can become one extra proficiency from a short list (there better be a choice, otherwise all PCs from the same family always have that skill); I'm thinking to make this more valuable, by allowing to gain skill expertise if you already have that skill proficiency
- some extra equipment: I'm undecided whether to keep this or remove

Ancestor in d20 is a feat, that can (and has to) be taken only at 1st level, and only 1 per PC. I think the best idea is to keep the d20 Ancestor feats as they are now, and each Human PC gets one. Ancestor feats that are inappropriate for 5e (e.g. those which break bounded accuracy) will either be adjusted or removed.

If this "package" is too weak overall, it might be considered to add some individual bonuses such as:

- bonus proficiencies on things other than skills
- still keep +1 to two ability scores of choice

 

[Li Shenron]

Preliminary thoughts: Classes

As I mentioned in the previous post, I'll focus only on the 4 main classes of the game, and assume that there is no multiclassing between them (characters do multiclass into Monk at a certain age, but we can ignore that for now).

Samurai

Heavy focus on martial features and abilities, which means plenty of feats, and optionally maneuvers. HD and weapon/armor proficiencies can stay the same as in d20.

Significantly more skilled than Fighters, and definitely not just physical skills. But 5e PCs are all more skilled than in 3e, so just one class skill proficiency is probably enough.

The only unique feature in the d20 Samurai was the Ancestral Daisho, but it had an XP cost to increase its benefits, so it wasn't strictly required to. I think the best mechanical representation for this is to create a subclass: at every level where you get a subclass benefit, your Ancestral Daisho improves. The XP cost is removed, but it improves by level so in a sense it's quite the same, without the sour taste of reducing your XP.

"Service classes" of the Samurai in d20 were mainly the Fighter, Barbarian and (spell-less) Ranger. My favourite idea is to break down these classes, and turn them into subclasses of the Samurai. Some class features of the 5e Fighter might simply become class feature of the Samurai, but for instance Rage can be very much unique to a Berserker subclass, which in turn might be restricted to characters of the Unicorn clan or the Hida and Matsu families.

Battlemaster and Warrior subclasses are totally appropriate also as Samurai subclasses. Battlemaster's maneuvers however, seem perfect to be restricted by Clan.

Monk

Basically the core D&D Monk was used in d20 Rokugan, only with some more flexibility in the choice of class features. Very likely we can use the 5e Monk as-is, since that is also largely based on the 3e Monk.

The current 5e Monk subclasses Way of the Open Hand and Way of the Elements are also quite appropriate.

d20 Rokugan had another Monk-like class called Inkyo, to represent a more supernatural type of monks. This might be turned into a further subclass (it's not very "elemental" but more "void-oriented", but another option is to merge it with the Way of the Elements).

Finally, I never quite sorted out whether the Sohei is or isn't a proper Rokugan class. In d20 Oriental Adventures it's said it doesn't exist in Rokugan, while the d20 Rokugan Campaign Setting mentions it as a type of Monk. I would be positive about having the Sohei as a Monk subclass, but definitely without the spells it has in Oriental Adventures.

Courtier

This class is a hybrid between a Rogue and an Aristocrat, minus the combat sides of both. The basic math of the class can be replicated trivially, but the real "meat" is skills and special Courtier abilities.

This class in d20 is a skillful as a Rogue, but didn't have physical skills in its class list. In D&DNext skills are largely individual rather than derived from class, while skillful classes (Rogue, Bard and Ranger) get 2-3 more skill proficiencies than anybody else, therefore this seems a must-have for the Courtier as well (with appropriate choices for the bonus skills list).

There is a handful of Courtier class abilities in d20 but not all of them are essential: IMHO the most interesting one is Gossip, the others may be retained or replaced by something else. But besides those common Courtier abilities, there are also the "special Courtier abilities" which are chosen, i.e. different to each character, and which ones you can choose depend on your Clan. Because of this 'selectability' and the restrictions by Clan, special Courtier abilities play a role similar to the Samurai's technique feats (and eventually maneuvers) and the Shugenja's spells, which makes them a must-have. Luckily, because they are a separate system, they don't need to be converted, they can just stay as they are, provided that whichever of them breaks bounded accuracy is removed or adjusted.

As for subclasses, we might try to fit a Bard~ish idea under the Courtier, particularly for providing inspiration benefits, but it should probably not be magical.

Shugenja

Definitely the most difficult of the bunch to adapt to D&DNext! The class math and general structure is not an issue, but the spellcasting system is.

I am actually not familiar with the magic system of L5R other than its d20 adaptation, which according to many was not appropriate enough: the Shugenja was basically a 3e Sorcerer (daily slots, fixed number of spells known, no need for preparation), but there is no such character in 5e where every spellcaster has to prepare spells in advance, so it's probably even further away from how magic is supposed to work in Rokugan.

Nevertheless, it might work to just use 5e spellcasting system. The alternative is to port a system from any other edition of L5R, but this may or may not fit with the rest of D&DNext.

Prestige Classes

May not be needed for a "basic Rokugan campaign", but at least let's take a look forward...

There were lots of PrCls in Rokugan d20, the majority of which represented training with a specific group, and were therefore restricted by Clan, but usually there is a much stronger emphasis on the narrative of becoming a member of a PrCl compared to standard D&D.

Anyway, older d20 Rokugan prestige classes can be converted into D&DNext in at least the following ways:

- feat chains
- subclasses
- 3e-like prestige classes

These are the main implication of choosing each approach:

As a feat chain, this means
(a) it's generally available to all base classes, but some classes may advance faster than others
(b) minimum character level to enter is 4th
(c) benefits are scattered over distant (fixed) levels
(d) possible to pace advancement only more slowly
(e) possible to take multiple prestige classes

As a subclass, this means
(a) it's available only to one class
(b) it has to start when the subclass starts, usually 2nd or 3rd, not later
(c) benefits are scattered over distant (fixed) levels
(d) advancement is automatic, pace is fixed
(e) only one prestige class is possible (unless the option of mixing multiple subclasses is allowed)

As a 3e-like prestige class, this means
(a) it's available to whatever classes have access to the requirements
(b) there might be an implied minimum level, but otherwise can start at any level
(c) benefits are achieved quickly, at consecutive levels
(d) advancement is freely paced
(e) possible to take multiple prestige classes

I really think that there is no one-size-fits-all, therefore all three approaches should be used, the best one depends on the specific prestige class to convert.

 

[Li Shenron]

Preliminary thoughts: Combat

In Rokugan, combat (particularly duels) is usually expected to be fast and deadly, while D&D generally treats combat as a matter of attrition, gradually decreasing each other's HP.

I don't want to change how combat works in D&DNext. I like its simplicity, I like how "theatre of the mind" (i.e. without keeping precise track of distances and positions) is much easier than in the previous 2 editions, and anyway changing combat rules can have lots of consequences on balance and other areas of the games. Overall, a lot of work I'd rather avoid.

However, D&DNext description of combat, damage and HP, seems to suggest to treat most damage as minor physical wounds and wear, or even fatigue, and to think that only strikes that drop to negative HP are "real" wounds. This might play in our favor, because if we adhere to this descriptive practice, we can imagine a fight in Rokugan being a matter of parrying and dodging, except that when an attack "hits" this means your defenses are worn down and you get closer to receiving that one hit that defeats you (or simply, the probability of losing the fight increases every time an attack gets "too close", even if we still describe a "hit" as a parry, dodge or perhaps cutting your hair and clothes).

Duels

While regular 5e combat rules might be fine after all when fighting monsters, duels are a little bit different.

In d20 Rokugan there are Iaijutsu rules for duels. They are supposed to significantly increase the chance of winning with your first strike. They don't always work, because even tho you get a huge damage bonus on your first strike (if you succeed at striking first!), the opponent's HP are normally more. It's very similar in magnitude to Sneak Attack, except that how much extra damage you deal depends on a skill check, and Sneak Attack definitely doesn't always kill an opponent in one blow.

Of course not every duel is to the death, many are to first blood, in which case the exact damage dealt is irrelevant, so the d20 Iaijutsu rules are still ok.

Still, I think it's worth starting off by trying to use the old d20 Iaijutsu rules for duels also in case of duels to the death, even if that means a few rounds of combat are usually due, which isn't exactly very "quick and deadly".

These are the needed adjustments:

- The d20 Iaijutsu rules are governed by a skill called Iaijutsu Focus. At first thought, we might just add this to the list of 5e skills, however the bounded accuracy of skills proficiency bonus might be a problem. This skill is not used for challenges against a DC, so what really matters is the relative difference with your opponent's bonus (i.e. your chance of striking first). There are supposed to be quite large differences between someone completely untrained and a master, and the bounded accuracy probably doesn't capture that. Still, there might be other ways to represent mastery, e.g. a Samurai subclass Iaijutsu Master (converted from the original PrCl) might get an additional bonus to Iaijutsu Focus, a feat might grant Expertise to Iaijutsu Focus, and so on.

- The absolute result of your Iaijutsu Focus determines the bonus damage (only if you do win the contest of striking first). In d20 Rokugan this damage can also be used in a regular fight if you catch an opponent flat-footed, so it has to be decided if we want something similar in 5e or if we only want this to work in a duel. In any case, the actual table for extra damage based on your skill check result certainly needs adjustment since the skill bonus/rank doesn't grow as high as in d20.

 

[Li Shenron]

Preliminary thoughts: feats in general

Some more thoughts about how to deal with feats, especially since d20 Rokugan has an immense material when it comes to feats... it would be great to be able to use them without too much work! But very clearly, feats in D&DNext are bigger, something between twice and three times the size of 3e feats.

One existing problem I already mentioned, is how to represent Void Use, which in d20 Rokugan is a large feats tree: the first feat gives you the ability to spend void points for basic uses, after that there's plenty of feats granting additional uses of void points or increased benefits for previous uses. Every character can potentially acquire these feats.

Another problem is martial arts feats. I would rather have these exclusives to Monks, so that Samurai have techniques and maneuvers, Shugenja have spells, Courtiers have special courtier abilities, and Monks have martial arts. All of these can sometimes be common and some other times Clan-exclusive.

It might be possible to salvage all these d20 feats (well, not all since we still have to purge the lists from feats which break bounded accuracy), so here's a very simple idea...

...how about if, every time a 5e Rokugan character can acquire a feat or +1 in two ability scores, she can either:

- gain a 5e feat
- gain +1 in two ability scores
- gain two 3e feats
- gain +1 in one ability score and one 3e feat

This would allow to both use all 5e feats (excluding those thematically inappropriate to Rokugan) and all 3e Rokugan feats (excluding those mechanically inappropriate to 5e).

I.e., we don't really need to worry about converting lots of smaller d20 feats into larger 5e feats!

 

[Li Shenron]

Preliminary thoughts: Skills and Tools

No reason to rush this task, since the whole 5e list of Skills and (especially) Tools might not be finalized yet.

What is almost certainly final, is that skills in 5e are less and therefore slightly more encompassing than 3e skills, that the choice of skills and tools proficiencies is almost completely free (no cross-class skills at double price, no restricted skills), that the general philosophy of 5e is "everybody can try anything" (but some exceptions might happen with tools), and that the total bonus is limited to +12 if you can get both proficiency and expertise.

As a matter of fact, skills in Rokugan are important distinguishing features of the characters. I don't think the idea "you can try anyway" works well in Rokugan... you don't "try" in Rokugan, you are supposed to excel at something or have someone else do it for you. It's not a problem for skills used in adventuring (such as Perception, Athletics but also Knowledge), it's really only a problem with skills used in society, particularly in court. For those, I think the idea of expertise should be leveraged, and so it should be available via feats or (sub)class features.

Besides that, Rokugan needs more skills than the core list. On the top of my head, skills such as Craft (for artistic purposes) and Geography (possibly region-specific) would be nice to have, as well as some Rokugan-unique skills such as Battle, Tea Ceremony and the already mentioned Iaijutsu Focus. Knowledge skills are a bit messy, there were too many in Rokugan, often overlapping (even Arcana and Religion are in a sense the same thing in Rokugan).

Finally, some skills/tools such as Deception, Animal Handling, Sleight of Hand, Thieve's Tools and even Medicine are sometimes (if not always) dishonorable, and should be either removed from the list, or care should be taken by the players to use them only honorably.

 

[Li Shenron]

First sanity checks with D&D 5e Basic

Now that the Basic Rules are finalized and released, I can get a better opinion on what I need to do for this conversion work!

Here are a couple of topics with easy resolution: the skills list and the human race. These are still preliminary thoughts, nothing set in stone yet.

Design mark #6: remove dishonorable (sub)skills from the list

The current list of skills in Basic is:

Acrobatics
Animal Handling
Arcana
Athletics
Deception*
History
Insight
Intimidation
Investigation
Medicine*
Nature
Perception
Performance
Persuasion
Religion
Sleight of Hand*
Stealth
Survival

Skills marked with * are most of the time dishonorable, but I am reluctant to remove them from the list. At least Scorpion characters are very likely to train in Deception anyway, and Sleight of Hand may have some uses that aren't dishonorable after all. Maybe the best is just leave all these, and just adjudicate (dis)honorable usage only during the action itself.

Unlike the skills list, the current list of tools may not be complete. Anyway many of these are going to be mostly dishonorable: disguise kit, forgery kit, poisoner's kit, thieves' tools certainly are, as well as all artisan's tools related to low crafts, but even these could be honorable when used "artistically". At least the following should be allowed:

Alchemist's supplies
Calligrapher's supplies
Cartographer's tools
Cook's utensils
Glassblower's tools
Jeweler's tools
Painter's tools
Potter's tools
Smith's tools
Weaver's tools
Woodcarver's tools
Gaming sets
Herbalism kit*
Musical instruments
Navigator's tools*
Vehicles

* not completely sure but probably ok

Design mark #7: identify what iconic Rokugani skills are missing from the list

Some skills that I'm currently thinking about adding to the current list:

Battle
Bonsai*
Calligraphy*
Concentration
Etiquette (may fold with Persuasion)
Geography
Iaijutsu
Ikebana
Landscape Gardening*
Maho
Origami
Tea Ceremony*

Skills marked with * might be more appropriate as tools.

I have some issues with knowledge skills, because in Rokugan magic is religion and it's about the spirits of nature, hence Arcana ~ Nature ~ Religion aren't that different from each other.

Design mark #8: no backgrounds, freely pick your trait+skills+tools

Backgrounds in their final form grant:

- one feature
- two skills
- two tools or languages
- equipment package
- two traits, one ideal, one bond, one flaw (all optional)

The structure is sound, and will be used by characters in my conversion to pick their choices freely, instead of using predefined backgrounds.

The only problem is that at the moment most existing backgrounds features are inappropriate, because they rely on social relationships and ranks. These cannot be applied as written in Rokugan, where caste, clan, schools and family already establish relationships. It makes no sense for someone to have the "Rustic Hospitality" feature, when every PC is a noble and the common folk are supposed to obey to all nobles, or the "Military Rank" feature which is supposedly automatic to all samurai. "Researcher" is the only feature that is fine as-is, all the others at the very least need serious adjustments. I'll have to wait for the PHB to see if I have a decent amount of backgrounds to choose from.

So instead of adding these three choices, I will replace the +1 to all abilities with the choice of Clan, Family and Ancestor.

Good news that we got variant human traits in Basic, meaning that I can already check my preliminary idea against such option:

- +1 to two different ability scores
- one extra class skill
- one feat

Clan and honor in my proposal don't weight anything with regard to balance, so my original suggestion could have summed up as:

- +1 to two different ability scores
- one extra class skill or expertise
- one ancestor feat (imported from 3e)
- one more non-skill proficiency

We are remarkably close! The ancestor feat is worth half a 5e feat at best, so if by "non-skill proficiency" we actually mean a weapon, armor, tool or language proficiency (but not a saving throw, which is more valuable), then we might actually have a winner.

Note that both the skill and non-skill bonus proficiencies should be restricted by family which is an important narrative choice in Rokugan, but the choice of family should provide a short list, not a fixed choice.

Example: you choose the Kobayashi (made up) war-loving family for your PC, and this grants you the following:

- proficiency in either the Athletics, Battle, History or Intimidation skill (or expertise if you are already proficient)
- proficiency in either one weapon of choice, one armor type or shield, or land vehicles

 

[Sanglorian]

Some skills that I'm currently thinking about adding to the current list:

Battle
Bonsai*
Calligraphy*
Concentration
Etiquette (may fold with Persuasion)
Geography
Iaijutsu
Ikebana
Landscape Gardening*
Maho
Origami
Tea Ceremony*

Skills marked with * might be more appropriate as tools.

I think Tool proficiency: Bonsai, landscaping rake, tea, origami paper are very cool, and work much better than skills. There's already a calligrapher's kit, from memory, so that won't require a skill.

Backgrounds in their final form grant:

- one feature
- two skills
- two tools or languages
- equipment package
- two traits, one ideal, one bond, one flaw (all optional)

The structure is sound, and will be used by characters in my conversion to pick their choices freely, instead of using predefined backgrounds.

The only problem is that at the moment most existing backgrounds features are inappropriate, because they rely on social relationships and ranks. These cannot be applied as written in Rokugan, where caste, clan, schools and family already establish relationships. It makes no sense for someone to have the "Rustic Hospitality" feature, when every PC is a noble and the common folk are supposed to obey to all nobles, or the "Military Rank" feature which is supposedly automatic to all samurai. "Researcher" is the only feature that is fine as-is, all the others at the very least need serious adjustments. I'll have to wait for the PHB to see if I have a decent amount of backgrounds to choose from.

What about having clans or families as backgrounds? That would fit well with different clans having different personality traits, flaws, bonds and ideals. And you mentioned that ancestor feats are self-contained - so why not make them ancestor features instead?

 

[Li Shenron]

Thanks a lot for the comments

I think Tool proficiency: Bonsai, landscaping rake, tea, origami paper are very cool, and work much better than skills. There's already a calligrapher's kit, from memory, so that won't require a skill.

I am uncertain about whether they'd better be skills or tools. My guess is that you definitely need some tool for tending to a bonsai, landscaping a garden, or prepare tea. OTOH you don't need any tool for ikebana as far as I know. You may create origami from origami paper, but you can also improvise them with normal paper, and calligraphy may apply also to writing with carbon, chalk or even on the sand.

These are just thoughts. Truth is that it doesn't matter that much whether technically they are skills or tools, also because these are skills that are probably rarely used (except tea ceremony which has its own pseudo-magic effects in-game). But somehow I'll have to make a choice for presenting these options to the players.

What about having clans or families as backgrounds? That would fit well with different clans having different personality traits, flaws, bonds and ideals. And you mentioned that ancestor feats are self-contained - so why not make them ancestor features instead?

I thought about the first idea, but it won't work IMO because a background is a fixed set of proficiencies. I don't want all Cranes or all Daidoji family members to end up with exactly the same skills proficiencies. Well, they still get 2 skills from class, but my point is that I don't want everyone in a clan - which is thousands of people at least - to have those 2 skills + 2 tools or languages. I think it's too restrictive.

About ancestor feats, if you mean for them to replace the background's "feature", it's probably ok even tho typically the latter is a downtime benefit. As a general principle, I'd like to try and keep things as close as possible to 5e default rules, so if I find enough backgrounds with applicable feature, I would prefer to re-use those.