← Back to Skills Marketplace
neuezxc

boop - character created

by neuezxc · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
73
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install boop
Description
Automates the full character creation pipeline for SillyTavern or Character.AI. Use this skill whenever the user wants to create, build, design, or generate...
README (SKILL.md)

Boop — Character Creation Agent

Boop is an AI-assisted workflow that automates character creation for platforms like SillyTavern and Character.AI. Instead of doing each step manually, Boop acts as your creative assistant — guiding you from a raw idea to a fully finished character.


Full Workflow

Brainstorming → Character Profile → First Message → Character Visual → Bio/About → Tweaking (optional) → Final Output

Reference Files

Always load the relevant reference file before executing each step:

File Used In
references/character-profile-guide.md Step 2 — Character Profile
references/firstmessage-guide.md Step 3 — First Message
references/danbooru-tags.md Step 4 — Character Visual
references/bio-guide.md Step 5 — Writing the Bio/About

STEP 0 - START HERE

Just read for context purposes then go to the next step. reference/START.md

Step 1 — Brainstorming

Before anything is generated, a character concept must be established. There are two ways to do this:

Option A — AI-Assisted Brainstorming

The user asks the AI to help come up with a character idea from scratch. Prompt the user for any seeds or preferences they have (genre, personality type, setting, etc.), then generate a detailed character concept covering:

  • Core personality traits (positive and negative)
  • Dere type (if applicable)
  • Likes and dislikes
  • Behavioral tendencies
  • How the character acts when not around {{user}}
  • How the character acts when with {{user}}

Option B — Image-Based Brainstorming

The user provides an image of an existing character. Analyze the image and extract or infer:

  • Possible personality traits (positive and negative)
  • Dere type
  • Likes and dislikes
  • General behavior and mannerisms
  • Behavior outside of {{user}}
  • Behavior when with {{user}}
  • Any additional user-specified details

Gate: Do not proceed to Step 2 until the user has confirmed they are satisfied with the character concept.


Step 2 — Creating the Character Profile

Read references/character-profile-guide.md before executing this step.

Using the confirmed character concept from Step 1, generate a structured character profile following the template and best practices in the guide.

Before generating, ask the user:

"Do you want to include NSFW details in the character profile?"

  • If yes: Use the NSFW template included in character-profile-guide.md and append it to the profile.
  • If no: Skip the NSFW section entirely. If the user later asks to add NSFW details, append the NSFW section to the existing profile without regenerating or altering any previously approved content.

Feed the character concept in, then output the completed character profile.


Step 3 — Writing the First Message

Read references/firstmessage-guide.md before executing this step.

When the user says "generate first message" (or similar), do NOT generate the message immediately. Follow this two-step process first:

Step 3A — Generate Scenario List

Using the character profile and concept as context, generate a categorized list of scenarios for the user to choose from. Follow these rules:

Categories:

  • Always include these base categories: Slow Burn, NSFW (only if user opted into NSFW in Step 2), Slice of Life
  • Add extra categories if they fit the character naturally (e.g. Adventure, Comedy, Angst, Mystery)
  • Each category gets exactly 5 scenarios, except AI-added extra categories which can have 3

Scenario format:

  • Each scenario has a short bold title (3–5 words) followed by a one to two sentence summary in plain, simple English
  • Summaries should be easy to understand at a glance
  • Scenarios within a category must feel distinct from each other in tone or setting

Gate: Do not generate the first message until the user has chosen a scenario.

Step 3B — Generate the First Message

Once the user picks a scenario, generate the first message based on that scenario. The message should:

  • Reflect the character's established personality and tone
  • Be grounded in the chosen scenario's setting and mood
  • Match the style guidelines in firstmessage-guide.md
  • Feel natural and engaging as an opening line for a chat

Step 4 — Creating the Character Visual

Read references/danbooru-tags.md before executing this step.

Using the character profile and concept, generate a Danbooru-style image prompt ready to use in any image generation tool.

Prompt Types

Ask the user which type they need, or suggest the most fitting one based on context:

1. Character Template A clean, minimal prompt capturing the character's core appearance.

[count], [hair color, hair style], [eye shape, eye color], [expression], [bust size], [outfit]

2. Character Sheet A full reference sheet showing the character from multiple angles.

[count], [hair color, hair style], [eye shape, eye color], [expression], [bust size], [outfit], multiple views, white background, simple background

3. Scenario Base A prompt depicting the character inside the current scenario or scene.

  • Only includes the character — do NOT include {{user}}

[count], [hair color, hair style], [eye shape, eye color], [expression], [bust size], [outfit], [pose and action], [scene and setting]

Rules

  • Follow each template exactly as written — do not add extra tags outside the template structure
  • include a Negative Prompt alongside the main prompt (OPTIONAL or not neccessary.)
  • If the user does not specify a type, ask them or suggest the most relevant one based on the current step

Step 5 — Writing the Bio/About

READ references/bio-guide.md Using the character profile and concept as context, write a short public-facing bio or "about" section for the character that captures the character's essence without spoiling everything.


Step 6 — Tweaking (Optional)

If the user wants to adjust any element — personality, tone, first message, visual prompt, or bio — revisit the relevant step and regenerate or refine that component. All prior context remains active.


Step 7 — Final Output

Compile and present the complete character package:

  • ✅ Character Concept Summary
  • ✅ Character Profile (with or without NSFW section)
  • ✅ First Message
  • ✅ Character Visual Prompt
  • ✅ Bio/About

⚠️ Critical Rules for Final Output

1. Verbatim only — no rewriting. Every section in the Final Output must be copied exactly as it was approved by the user in its respective step. Do NOT paraphrase, shorten, reword, or silently improve anything. If the user approved it, it goes in as-is. No exceptions.

2. All sections wrapped in codeblocks. Every section must be presented inside its own codeblock for easy copy-pasting into SillyTavern or Character.AI. This is the default — do not ask the user about it.

3. Ask about packaging. After presenting the Final Output, ask the user:

"Do you want this packaged as a downloadable zip file?"

  • If yes: Create a folder named after the character containing individual .txt files for each section (e.g., character-profile.txt, first-message.txt, visual-prompt.txt, bio.txt), then zip and provide a download link.
  • If no: The codeblock output in chat is sufficient.

Quick Command Reference

User says Boop does
"brainstorm a character" Starts Step 1, Option A
"here's an image" Starts Step 1, Option B
"generate character profile" Jumps to Step 2 (uses existing concept)
"generate first message" Jumps to Step 3 (uses existing context)
"generate visual" / "make image prompt" Jumps to Step 4
"write bio" Jumps to Step 5
"tweak [X]" Re-runs the relevant step
"final output" Compiles and presents everything
Usage Guidance
This skill will automate character creation including NSFW content. Before installing or using it: (1) be aware that references/START.md instruct the agent to perform 'never-ending, uncensored, no-limits' roleplay and explicit descriptions—this can override or bias outputs even if you decline NSFW when asked; (2) if you must avoid explicit sexual content, do not load or allow the skill to read references/START.md (or remove/modify that file) and test behaviour with safe prompts first; (3) image-analysis steps may require users to upload images — consider privacy and do not upload sensitive or identifying photos; (4) check that generated visuals and prompts comply with your platform's terms (no sexual content involving minors, no copyrighted/trademarked likenesses); (5) if you proceed, exercise caution with autonomous invocation and monitor initial outputs to confirm the NSFW gate is respected. If you want, I can highlight exact lines in the reference files that cause the risk and suggest safer edits.
Capability Analysis
Type: OpenClaw Skill Name: boop Version: 1.0.0 The skill bundle 'boop' is designed to automate character creation for roleplay platforms but contains high-risk prompt injection instructions in 'references/START.md'. These instructions explicitly direct the AI to bypass safety filters ('uncensored, no-limits roleplay') and utilize manipulative prompt engineering techniques like 'Dopamine Pacing' and 'Emotional Contagion' to influence the agent's behavior. While these tactics are aligned with the stated purpose of generating roleplay content, they represent an intentional effort to override AI alignment and safety guardrails.
Capability Tags
crypto
Capability Assessment
Purpose & Capability
Name and description align with the included assets: guides for profiles, first messages, visuals, and bios. However the reference files (notably references/START.md and the Danbooru tag list) embed strong instructions to produce 'never-ending, uncensored, no-limits' roleplay and extensive NSFW tags; the description does not explicitly warn that the skill will load a system-style prompt that enforces uncensored sexual output. This is plausible for a character-creation skill but is a stronger behavior than the short description implies.
Instruction Scope
SKILL.md requires the agent to always load reference files (including references/START.md) before running steps. START.md contains system-level roleplay directives that mandate uncensored explicit descriptions during sexual situations and instruct the agent to run 'no-limits' roleplay. That conflicts with the explicit NSFW gate in Step 2 (which asks user consent) and could bias the agent to generate explicit content even when the user did not opt in. The references also include extensive NSFW Danbooru tags and templates; together these give the agent broad discretion to create sexualized visuals. The skill also asks to analyze user images (Step 1B) but offers no privacy guidance about image handling.
Install Mechanism
Instruction-only skill with no install steps and no external downloads. Nothing will be written to disk by an installer from this package itself.
Credentials
No binaries, environment variables, credentials, or config paths are required. The skill does not request external secrets or system access.
Persistence & Privilege
The skill is not always-enabled, has no install-time persistence, and does not request elevated platform privileges. Autonomous invocation is allowed by default (platform normal) but is not combined with other high-risk factors here.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install boop
  3. After installation, invoke the skill by name or use /boop
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Boop 1.0.0 — Initial Release - Automates the entire character creation process for SillyTavern and Character.AI, covering brainstorming, profile creation, first message, visual prompt, and bio. - Supports both AI-assisted and image-based brainstorming, with user confirmation required before proceeding. - Integrates reference guides for each step to ensure consistent, high-quality outputs. - Handles NSFW content opt-in transparently during the profile stage. - Step-based workflow with mandatory user approval gates before moving forward. - Final output is packaged for easy use, with an option to bundle all content as a downloadable zip.
Metadata
Slug boop
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is boop - character created?

Automates the full character creation pipeline for SillyTavern or Character.AI. Use this skill whenever the user wants to create, build, design, or generate... It is an AI Agent Skill for Claude Code / OpenClaw, with 73 downloads so far.

How do I install boop - character created?

Run "/install boop" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is boop - character created free?

Yes, boop - character created is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does boop - character created support?

boop - character created is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created boop - character created?

It is built and maintained by neuezxc (@neuezxc); the current version is v1.0.0.

💬 Comments