google-agents-cli-scaffold
Project scaffolding, deployment configuration, and CI/CD setup for Google ADK agents.
character
Core patterns for all characters — home, location, relationships, inventory
Full skill instructions
"File is identity. Location is presence. Relationships are memory."
Characters are entities that exist in the world. Players, NPCs, companions, cats — all are characters.
The critical distinction:
player:
home: characters/don-hopkins/ # Where FILE lives (never moves)
location: pub/ # Where CHARACTER is (changes)
| Concept | Purpose | Changes? |
|---|---|---|
| home | Physical file path | NEVER |
| location | Current position in world | Runtime |
Why: Stability, safety, git-friendly diffs.
Characters own their own state. The CHARACTER.yml file is CANONICAL for:
| State | Stored In | Notes |
|---|---|---|
location | CHARACTER.yml | Where the character is in the world |
inventory | CHARACTER.yml | What the character carries |
gold | CHARACTER.yml | Character's resources |
sims_traits | CHARACTER.yml | Personality (Sims-style) |
mind_mirror | CHARACTER.yml | Personality (Leary-style) |
relationships | CHARACTER.yml | Who they know and feel about |
memories | CHARACTER.yml | What they remember |
ADVENTURE.yml (or other world state files) may MIRROR some state for convenience, but CHARACTER.yml is always the source of truth.
# CHARACTER.yml (CANONICAL — edit this first)
player:
location: coatroom/ # CANONICAL — character owns their location
inventory: [lamp] # CANONICAL — character owns their inventory
# ADVENTURE.yml (MIRROR — optional convenience copy)
player:
location: coatroom/ # Mirror of CHARACTER.yml
When updating state:
Why this matters:
characters/
don-hopkins/ # Directory name = character ID
CHARACTER.yml # The character file
CARD.yml # Optional: makes character a playable card
cookie-1.yml # Dispensed items
notes.yml # Personal stuff
Any character directory can have a CARD.yml sidecar to make the character card-playable:
# characters/don-hopkins/CARD.yml
card:
for: ./CHARACTER.yml
type: hero-story
tradition: "Pie menus, SimCity, constructionism"
advertisements:
SUMMON:
description: "Activate Don's design traditions"
See card skill for full pattern.
pub/
bartender.yml # Lightweight NPC
cat-cave/
terpie.yml # Full family as single files
stroopwafel.yml
kitten-myrcene.yml
Rule: Need junk storage? → directory/CHARACTER.yml. Just a character in a room? → character-name.yml.
"Does this character BELONG to a place, or VISIT places?"
| Type | Home | Examples |
|---|---|---|
| Belongs | Room directory | pub/bartender.yml, maze/skeleton.yml |
| Visits | Own directory/repo | characters/don.yml, github:user/char.yml |
Key = other entity ID. From is implicit (file owner).
# In marieke.yml
relationships:
don-hopkins:
feeling: "A regular now. One of the good ones."
memories:
- "The day he sat with Myr for three hours"
hopes: "I hope he keeps coming back."
self: # Private inner data!
identity: "Third generation. This is who I am."
fears: "That I'm not enough."
mantra: "The cats need me. I need them."
Relationships can point to anything:
don-hopkins)brass-lamp)pub/, maze/)acme-corporation)self) — private storage!| Level | Score | Effect |
|---|---|---|
| Stranger | 0-15 | -10% success, 50% effects |
| Familiar | 41-60 | +10% success, 100% effects |
| Friend | 61-80 | +20% success, greetings |
| Soulmate | 91-100 | +50% success, psychic link |
Objects stay in their home. Picking up = references, not file moves.
# Don picks up kitten:
# In don-hopkins.yml
inventory:
- pub/cat-cave/kitten-myrcene.yml
# In kitten-myrcene.yml
location: characters/don-hopkins/inventory
# File didn't move. Location changed.
Reset: Snap objects back home: location = home.
Items in inventory can be OBJECTS or REFS:
| Type | Weight | Bulk | What It Is |
|---|---|---|---|
| Object | Yes | Yes | The actual item (lamp, sword, lunchbox) |
| Ref | No | No | Lightweight pointer to a prototype |
Refs are perfect for: catalogs, manuals, maps, guides — things you reference but don't physically carry.
inventory:
# Full object
- item: "Brass Lantern"
type: object
source: start/lamp.yml
weight: 2
fuel: 100
# Lightweight ref
- item: "ACME Catalog"
type: ref
prototype: street/lane-neverending/w1/acme-catalog.yml
annotations: ["circled portable hole", "margin notes on physics"]
Some objects dispense refs (like the ACME Catalog Dispenser):
When dropping a ref in a room:
[item-name].yml# kitchen/acme-catalog.yml — dropped instance
object:
name: "ACME Catalog"
type: instance
prototype: ../street/lane-neverending/w1/acme-catalog.yml
origin: "Torn from dispenser at 4 Lane Neverending"
dropped_by: "don-hopkins"
annotations: ["DO NOT ORDER Rocket Skates", "circled portable hole"]
inventory_capacity:
max_weight: 45 # Varies by character
max_bulk: 10
refs_free: true # Refs don't count!
Some objects clone full objects on pickup. Original stays, you get instance.
# pub/cookie-jar.yml
object:
id: cookie-jar
dispenser: true
instance_template: cookie
When picked up:
characters/don-hopkins/cookie-1.yml| Type | File? | Use For |
|---|---|---|
| Ephemeral | No | Quick transaction, one-line dialog |
| Persistent | Yes | Ongoing negotiation, relationship state |
Rule: Will this matter in 5 minutes? No → ephemeral. Yes → persistent.
Two systems:
| System | Scale | Origin |
|---|---|---|
| Sims Traits | 0-10 | The Sims 1 |
| Mind Mirror | 0-7 | Timothy Leary |
Distribution: Original Sims used 25 points across 5 traits. Good guideline.
| Command | Effect |
|---|---|
LOOK AT [char] | Description, visible state |
TALK TO [char] | Conversation based on relationship |
HELLO [char] | Initiate social interaction |
GOODBYE [char] | End interaction, dismiss ephemeral |
EXAMINE [char] | Deeper observation |
Characters can live in other repositories:
player:
home: github:donhopkins/characters/don.yml
location: pub/
Characters can be "at" a specific line in a file:
character:
name: schema-expert
home: characters/experts/schema-expert/
location: "@central/apps/insights/pyleela/brain/Schema.py:142"
# Currently examining line 142 of Schema.py
Location path syntax for code:
@repo/path/to/file.py — at a file@repo/path/to/file.py:42 — at specific line@repo/path/to/file.py:42-67 — examining line range@repo/path/dir/ — in a directory (room)> where is schema-expert?
schema-expert is at @central/apps/insights/pyleela/brain/Schema.py:142
examining the createSyntheticItemIfNeeded method
See room/ for directories as rooms and files as objects.
Form parties of expert characters to explore code together:
> summon drescher-expert, devops-expert, security-auditor
Party formed: [drescher-expert, devops-expert, security-auditor]
> party: review @central/tools/edgebox/scripts/ingest.py
drescher-expert: "The claim_task function is a schema pattern..."
devops-expert: "The PostgreSQL SKIP LOCKED is elegant..."
security-auditor: "I see proper parameter binding, good..."
> party: focus on lines 280-350
[Multi-perspective analysis of the code section]
Party Navigation:
> go to @central/apps/insights/pyleela/brain/
> enter Schema.py
> look at lines 140-180
> search for "synthetic"
> follow import ExtendedContext
> back
> party: gather here
Each party member brings their expertise. The LLM simulates all perspectives in a single call (see speed-of-light/).
Comments ARE the character's thoughts:
needs:
hunger: 3
# I could eat. But I'm comfortable here.
# The kittens are asleep on my lap.
Important distinction: When CHARACTER.yml specifies motivations, goals, or needs, these are declarative descriptions — they tell the LLM how to roleplay this character.
The LLM does not genuinely want to achieve these goals. It performs wanting them.
motivations:
- protect_the_kittens # Declared, not felt
- find_good_food # The LLM roleplays caring about this
- explore_new_places # No intrinsic drive — just character data
Why this matters:
See: representation-ethics/examples/aggregate-patterns.yml for the distinction between aggregate simulation (valid) and individual prediction (not valid).
K-line: MOTIVATION-ROLEPLAY
Project scaffolding, deployment configuration, and CI/CD setup for Google ADK agents.
Set up tracing, logging, and monitoring for deployed ADK agents across Cloud Trace, BigQuery, and third-party platforms.
Enterprise Azure infrastructure architect generating Bicep or Terraform from workload descriptions.
Plan and configure production-ready Azure Kubernetes Service clusters with Day-0 and Day-1 best practices.
Raw mechanical interfaces fusing Swiss typographic print with military terminal aesthetics. Rigid grids, extreme type scale contrast, utilitarian color, analog degradation effects. For data-heavy dashboards, portfolios, or editorial sites that need to feel like declassified blueprints.
Web search, scraping, extraction, crawling, and monitoring via ScrapeGraph AI CLI.
Skill for working with Firebase Hosting (Classic). Use this when you want to deploy static web apps, Single Page Apps (SPAs), or simple microservices. Do NOT use for Firebase App Hosting.
Deploy and manage web apps with Firebase App Hosting. Use this skill when deploying Next.js/Angular apps with backends.
Deploy applications and websites to Vercel. Use when the user requests deployment actions like "deploy my app", "deploy and give me the link", "push this live", or "create a preview deployment".
Build SEO-optimized pages at scale using templates, data, and proven playbook patterns.
Design and build isolated, reusable Convex backend components with clear boundaries and app-facing wrappers.
Deploy and manage projects on Vercel using token-based authentication. Use when working with Vercel CLI using access tokens rather than interactive login — e.g. "deploy to vercel", "set up vercel", "add environment variables to vercel".
Opus.ai: Revolutionize Your Web Experience
Build a no-code AI app in minutes.
An IDE for code migration from legacy to modern frameworks through coding agents.
Automate CGI animation in live-action scenes
Branded artistic QR-code concepts
Launch a website in seconds with AI.
Streamline Your Coding Experience with AI Code Helper
Convert any screenshot or design to clean code.