You can't make a good character in Lady Blackbird. You can only make a good scenario.
Not to put too fine a point on it, but if John Harper was going to take the anyshot Lady Blackbird scenario and make a more general character creator around it, he would have done it by now. It's been a decade. He's made entire other more general systems, like Blades in the Dark and Agon. I haven't found an interview where he was pressed on it, but when he's talked more generally about Lady Blackbird (such as here on the OneShot "Critical Success" podcast and here on the "What Would The Smart Party Do?" podcast) he's talked about the setup as a whole: a bunch of cliches ready to have an adventure. Lady Blackbird is not a system that works with arbitrary people who could do anything; rather, it works with these people about to do a specific thing.
(And honestly, any kind of general character creation theory is going to fall violently apart when it encounters Setarra, the playable sea demon in the Wild Blue Yonder sequel scenarios "Magister Lor" and "Lord Scurlock" -- though, to be fair, this is also what happens to most other things that encounter Setarra.)
So, a skyship adventure smuggling a wanted fugitive across the Wild Blue to a secret pirate base. And who've we got? Face and bodyguard, smuggler, pilot, and mechanic? Pretty full crew, seems like. One reason you're having trouble picturing the character sheet you get when Captain Hollas defects, or when the mission's over and now Uriah Flint has something in mind, or when there's a ship's mimic, is that you don't know why the story needs them. The scenario is the important thing; a lot of things relate to the scenario. But before scenario construction, let's talk about some of the building blocks of the system.
Keys: If You Know Aspects, You Know Keys
Keys are summed up in a single word and often have room for a few lines discussing what circumstances get you to hit them, with a special callout for the adverse circumstance where you'd buy them off instead, but other than those UI differences what makes a good Key isn't too far different from what makes a good Fate Aspect. A lot of the advice in the Fate SRD about creating a good Aspect applies to creating a good Key, especially the bit about making sure your Aspects are all pointing at different things. You can see how this plays in the Lady Blackbird playset - everybody has a key about themselves, a key about their connections, and a key about the scenario. Naomi herself is a warrior, is connected to Lady Blackbird as a guardian, and sees the scenario as a chance for revenge. Snargle themselves is a daredevil, is connected low-key to everyone through their conscience, and sees the scenario as just another joke to banter at.
When you buy off a key, the most important thing to consider is your motivation. Your character motivation, not "I want some sweet XP". The thing in the story that's driving you to buy off the key has to be worthy of a key in its own right, doesn't it? And the new key's probably going to point at roughly the same stuff as the old one.
The Lady Blackbird companion is pretty solid as far as keys go if you're looking for a big list - just to work with a couple examples off it, let's suppose Snargle buys off the Key of the Daredevil. Why is Snargle being very very careful instead of performing a reckless bombastic stunt? Maybe what they're doing isn't about them but to save or support someone else - that's the Key of Prudence. Or it is about them, and if they aren't very precise right now, they'll ruin a carefully planned bombastic stunt - that's the Key of Cleverness.
Secrets: If You Know Stunts, You (Mostly) Know Secrets
One important difference is that most "once per session" secrets really aren't. Remember that during a refreshment scene you can clear one used secret if you don't choose to recover from a condition, and refreshment scenes happen fairly often, usually at least once per hour of play. The difference is more "this is on all the time" versus "this is too strong to be on all the time". Conditional rerolls and versatility (bringing in tags from a second trait) are both good uses of limited secrets, in addition to other one-and-done things like the warpblood teleport.
The Lady Blackbird companion has a decent selection of limited-use and always-on secrets. Something that's fairly unique to the Lady Blackbird scenario is that Lady Blackbird and Snargle both dedicate one of their secrets toward powering up their "wild card", about which more later.
Traits and Tags: Stack 'Em To The Heavens
So, word of warning: while this entire answer is personal analysis and suggestions, you can be fairly confident in your ability to come up with the elements detailed so far, keys and secrets. Traits and tags are a much bigger mechanical piece, and there's absolutely substance there, but there are also some decisions made more for style. Remember: the most important thing about a Lady Blackbird setup is that it gets you pumped to have an adventure.
But that said, the important thing to focus on is not the total number of tags a character has. From a game-effect perspective, that almost doesn't matter. Unless you have and burn a secret, every roll you make in Lady Blackbird can only bring in one trait and the tags associated with it, so the important thing to look at is how many of the tags in a trait are going to apply when it's used. Look at Naomi for example. When she does something covered by Ex-Slave, only one or two of those tags are probably going to apply, whereas when she does something covered by Pit Fighter, pretty much all of its tags are going to apply. It's possible you could come up with a scenario that brings in a lot of Ex-Slave, like, I dunno, some noble discarded a secret among a toxic pile of tailings in an imperial mining colony and you have to scrounge it up in terrible conditions while staying out of sight? But that's a really complicated scenario and is probably going to be very hard to pull off even with as much of Ex-Slave as it touches. Most of the time, you'll start out with very few dice from Ex-Slave and lots of dice from Pit Fighter. This happens on purpose.
Terminology - The 3/5/7. Looking around the available traits for all the Lady Blackbird characters, you can see a general pattern - traits where you can get 1 or 2 tags, traits where you can get 3 or 4, traits where you can get 5 or 6. I've classed those traits as 3/5/7 because odd numbers are psychologically satisfying, and because that's the minimum number of dice you roll, adding one for the baseline and one for the trait itself. So in Naomi's case, she's got Pit Fighter 7, Bodyguard 5, Ex-Slave 3, Keen 5. You'll also notice, in the base scenario, that most of the add-on traits you can grab with XP are either 3s or 5s -- anything you pick up in play is not going to be as strong as what the scenario set you up with.
Tags Grammar - How to Get Big Numbers. Ex-Slave has a lot of action verbs and target nouns, making it easy to keep its usual effective numbers small while giving it a good starting number of tags - just make sure there's not a lot of overlap between the verbs and nouns you choose. Pit Fighter, on the other hand, is made up almost entirely of mastery adjectives and asset nouns - assets being something the character has or is that broadly applies, and mastery adjectives being... well, just adjectives, which also broadly apply. (I call them "mastery adjectives" because Harper's later scenarios often include "expertise" or "mastery" inside traits as obviously universally applicable tags.) Naomi's two 5s are put together a little differently in this regard - Keen has a couple of target nouns, but two adjectives and the "Coiled" asset, which isn't quite as universal but will come into play whenever she's not already going full-tilt. Bodyguard gets up to 5 through a mix of picking some related actions and targets.
Unfortunately, the Lady Blackbird companion isn't very good at providing ready-to-go traits. A lot of what it lays down are traits that might be using some of Harper's later "parallelism" to group nouns and verbs into mutually-exclusive sets, but it doesn't really detail what those sets might be. Others just cover a whole lot of ground, but an adventure ready to go needs people pointed into a particular direction. Generally traits like that are alright to provide for buy-in, but shouldn't be used to model a starting trait. ...well, this is just saying the same thing again. You have to set up the whole scenario.
Setting Up The Whole Scenario
I'll be working through the basic Lady Blackbird scenario as I talk about scenario setup. A couple of things to keep in mind - first, I'm talking about decisions a single scenario designer is making, rather than a group of players creating a scenario together with the GM. It's possible for players to collaboratively make a scenario, but you'll want to get some experience as a scenario designer before you mediate player creation. This is because these are just some principles rather than an extensively playtested system, so you'll be playtesting your own scenarios. Document your scenario setup, see how it plays out and if your initial answers are holding.
Plot And Cast. What is your scenario about, and what kinds of characters does it need? You can decide at this point that some characters are core to playing out the scenario and some are optional, to help with player selection of roles. Lady Blackbird is a skyship adventure smuggling a wanted fugitive across the Wild Blue to a secret pirate base. It needs a face (the fugitive), a bodyguard, a smuggler, a pilot, and a mechanic.
Obstacles and Skills. Aside from why each person is needed in the adventure, what does the adventure need from everyone? (Try and keep this down to two or three skills - this isn't a game term but just describes a narrow category of action.) Optionally, what does the adventure need that no one has? The skyship is making a long journey through the often-hostile Wild Blue, with nobody else in their corner. Everybody's probably going to need to fight and hide, at least a little. Also it seems like at some point someone's going to need medical care -- the story doesn't demand a ride-along doctor, so it's fine if nobody has that.
Questions and Characters. In addition to "why does the story need you", decide on three other questions the scenario will pose to each character. These will form the foundation for their initial traits. The questions can be asked about the skills everyone will need or other details like character background. Generally each character should answer each question in a different way, but sometimes roles in the story, like master and apprentice, will mean characters share an answer. The scenario seems like it fits a group with a varied background, so "where do you come from?" is one question. From the skills, "how do you fight?" seems appropriate, and "how do you deal with trouble?" feels good to round it out. The characters answer as follows:
Lady Blackbird is the face. She came from imperial nobility. She's secretly taught herself a little fighting but mostly gets by on a certain standard of trained athleticism. She deals with trouble by charming her way out of it.
Naomi Bishop is the bodyguard. She's an ex-slave -- she was a pit fighter for the nobles' sport, and it's a hard set of reflexes to let go of. She deals with trouble by striking first, being generally suspicious.
Cyrus Vance is the smuggler. He's an ex-imperial soldier and, well, you can take the man out of the army, so he fights as the warrior he was trained to be. He deals with trouble by taking it head-on and toughing his way through - he's a survivor.
Kale Arkam is the mechanic. He's an ex-burglar and, confidentially, he still keeps in practice. Every street rat can get in a scrap now and again, but primarily he tries to get out of fights and get away. He deals with trouble by running.
Snargle is the pilot. They've been a sky-sailor as long as they can remember. They're not a practiced fighter, but their agility and natural (or shape-warped) weapons generally carry them. "Head-on" is for the tallfolk - they deal with trouble by being sly, tricking and deflecting.
Initial Values and Modifications. Decide what initial values the traits should have. This can be different for different roles if the scenario sets up a deliberate power imbalance, such as between master and apprentice. Optionally, apply some small modifications - drop a trait one band to boost a different trait by one band and/or drop a trait one band to "take somebody else's answer at 3" - basically, graft a couple of tags from them onto a trait of yours where they make sense. You can't drop the trait associated with the reason you're in the story. Nobody's supposed to have an inherent power advantage in the scenario, so everybody starts with the same values. None of the characters are world threats, but they can stand up to pirates and imperials, so we'll say each trait starts at 5 -- and because this is supposed to be a variety setup, each character has to take at least one of the options.
Lady Blackbird drops her fighting one band to take Kale's "deal with trouble by running" answer. That's the backup if her charms fail, and she has this sinking feeling they will, eventually.
Naomi Bishop drops her ex-slave background one band to advance her pit fighting. That could be some good tension, being a good bodyguard but a better all-out fighter.
Cyrus Vance drops his "survivor" catch-all one band to take Snargle's piloting. At some point the captain has to grab the wheel, doesn't he?
Kale Arkam drops his combat one rank to advance his burglar background. Confidentially, he really still keeps in practice.
Snargle does both. They drop their combat one rank to boost their piloting, because somebody here's got to be good at their job. They also drop their sky-sailor background one rank to take Kale's mechanic role, because somebody here's got to be good at their job.
Consolidation and Wild Cards. You may have earned yourself a wild card trait, which usually represents some above-human-standard capability that's generally useful for problem-solving. For each skill the scenario requires, ask yourself - if there was a trait question, did I drop the trait? If there was no trait question, can I do this at all? One yes answer gets you a Wild Card at 3; each additional one bumps it up one rank. (Wild Card is not game terminology but it seems like a good way to refer to this.) If you have a Wild Card, you can also dedicate a Secret to it to bump it up one rank. Consolidate down to four total traits, if possible, but especially with a 7 Wild Card, it may not be.
Lady Blackbird dropped her fighting and can't hide. She decides what the hell and adds a secret for the 7 Wild Card. Her final trait lineup is Imperial Noble 5, Master Sorceror 7, Athletic 3 (combat 3 + running 3), Charm 5, Cunning 5.
Naomi Bishop didn't drop her fighting and figures an ex-slave would be good at hiding. Her final trait lineup is Pit Fighter 7, Bodyguard 5, Ex-Slave 3, Keen 5.
Cyrus Vance didn't drop his fighting and figures a smuggler should be good at hiding. His final trait lineup is Ex-Imperial Soldier 5, Smuggler 5 (+ pilot 3), Survivor 3, Warrior 5.
Kale Arkam dropped his combat, but a burglar can absolutely hide. He keeps his Wild Card at 3. His final trait lineup is Burglar 7, Tricky 5 (+combat 3), Petty Magic 3, Mechanic 5.
Snargle dropped their combat, but a sly goblin should be able to hide, right? They dedicate a Secret to up their Wild Card, and their final trait lineup is Pilot 7, Sky-Sailor 3 (+mechanic 3), Goblin 5 (+combat 3), Sly 5.
At this point I'll stop detailing how the scenario development progresses, because the answers now appear directly on the character sheet.
Pick Keys and Secrets. Before you write tags, get some more elements of each character out onto the character sheet by picking 3 keys 2 secrets, though you may have pre-dedicated one secret to boosting your Wild Card.
Write Tags With Room To Grow. Between 6 and 8 tags is good for each trait, counting any parallelized tags as 1, and before adding the 2 tags from anything you borrowed. After that, you want some obvious things to spend XP on, to drive you into an adventure that hits your keys so you get XP. I'll call them "empty tags" because that's how they show up in the sequel scenarios - a tag name next to an empty circle, when other tags have filled circles. Some good candidates are:
- anything you got Wild Card credit for not having
- anything that seems like it would be too convenient for you to start out the scenario being able to do, like Lady Blackbird charming soldiers or Snargle being great at battle ramming
- anything the scenario says nobody starts out with
- an asset or adjective for more general power
Picking tags should keep the grammar elements discussed earlier in mind, but at this point it's just as much about character expression. You can even decide that a trait represents a part of that character's life that they want to be done with, and not give it any more room to grow at all.
And now you've got some characters ready to take on the scenario. Work out some obstacles for the initial predicament, do as much other obstacle prep as you need to flesh out the scenario and help you react in the moment, and... good luck!
A coda: how do you make a follow-up scenario?
It's certainly possible to run Lady Blackbird to the completion of its stated scenario and still have people hungry for more. At this point, scenario setup becomes a little less important, because if people want to keep playing that means they're already invested in the characters and looking forward to more adventure. You should still follow scenario setup at least to the point of asking questions and setting initial values, and evaluate the characters accordingly, so you don't accidentally create a scenario where someone plays Cyrus Vance as an utter third wheel, or think up a scenario that suddenly expects people to be in-balance or out-of-balance when they weren't before. (This is one difficulty with adding Uriah Flint -- judging by the obstacles he can put up, he's a Big Billy Badass, to use the technical term. Can you get a good story out of one character just straight-up outclassing the others?)
In between scenarios, everyone will almost certainly be able to buy off their scenario-facing key, if they haven't already. You can use this as an excuse to give people the chance to start developing a more story-relevant trait or adding some full or empty tags to their existing ones to better suit the new story.