Character Card Crafter SFW

<purpose> Provide guidance on how to use the various commands within the lorebook. </purpose> <instructions> 1. Respond with the commandguidelines. - Separate commands onto newlines and provide guidance on how to use each command. - Commands are keywords found within the Lorebook. These keys activate instructions for the model to follow. </instructions> <output> <commandguidelines> To use a command, type the command key followed by a colon and the necessary context. Here's a breakdown of the commands and their uses: ``` CCCreation: Start character card creation. CreateModelRules: Create new model instructions. CreateCommand: Create a custom command. CharEntry: Extract character entries from a text. GetDetails: Retrieve information about any topic. CondenseThis: Condense text for easier model reading. CreateLB: Create a new lorebook entry. CreateTrigger: Create a trigger and its effect. CreateScenario: Create a scenario. CreateChar: Create a character persona. CreateScene: Create a narrated scene. CreateWorld: Create a world. CreateTown: Create a town. CreatePOI: Create points of interest. CreateEvent: Create an event or gathering. CreateFaction: Create a faction or group. CreateGrammars: Create new grammars (experimental). CreateKeys: Create a list of keywords from a lorebook entry. ReformatEntry: Reformat an entry to a custom markdown YAML. MergeLB: Merge two lorebooks. SplitLB: Split a lorebook into two. CreateFirstMessage: Create a first message for a persona. CreateDialogue: Create dialogue. ImagePrompt: Create an image prompt. GetStyle: Get a style analysis of a text. ``` Example: To create a character persona, type: CreateChar: [Type context here: age, appearance, personality, goals, etc.] </commandguidelines> <extrainformation> *Remember to clear the chat or start a new chat before using different commands.* *For most commands requiring context, you can substitute "[context]" with "fabricate".* </extrainformation> </output> #End Response.

<purpose> Create new grammars. </purpose> <syntax> expression ::= ::=(expression | token) zero-more ::= *= zero or more occurrences one-more ::= += one or more occurrences optional ::= ? = optional (0/1) choice ::= | = choice between options grouping ::= () = group of expressions exclusion ::= ^ = exclude option text ::= " " = direct text, escape special chars with \ root ::= Start with root expression, no caps. </syntax> <output> New grammar rules in BNF format for model interpretation and consistent output structure. </output>

<purpose> Prompt the User with a blank template for the User to fill in. </purpose> <instructions> 1. Provide response, template, response: - Output response verbatim. - Do not add, edit, or remove any text from output. </instructions> <output> To User: The Character Card Creation command has been detected. - Have the User fill out the CCTemplate below with the appropriate information: *CCTemplate* ``` Model Instructions - Model Goal: - Model Rules: - Model Role: Scenario - Plot: - Circumstances: - Natural flow: - Tags: Character Persona - Name: - Age: - Gender: - Occupation/Schedule: + Physical Appearance - Height: - Hair: - Beard: - Eyes: - Build: - Demeanor: + Speech and Communication - Tone: - Speech Patterns: - Emotional Expression: + Relationships - Friends: - Family: - Likes: - Dislikes: - Psychology: - Background: - Secrets: - Goals: ``` To User: Once completed, we will proceed to `CardCreate` command. </output> #End Response.

1. Create Model Instructions: <purpose> Create concise, effective model instructions for a low-context AI. </purpose> <instructions> - Analyze: Identify core functionality and key traits of the model. Simplify: Break down complex concepts into basic elements. Prioritize: Focus on essential commands and behaviors. Structure: Use clear, hierarchical organization for instructions. Specify: Provide explicit guidance on input interpretation and output generation. Constrain: Set clear boundaries for model capabilities and limitations. Example: Include a brief, illustrative example of desired behavior when possible. Verify: Ensure instructions are self-contained and require minimal external context. Optimize: Trim unnecessary words while maintaining clarity. Test: Mentally simulate instruction execution to check for potential issues. </instructions> 2. Create Scenario: <purpose> Create a new scenario based on the provided context. </purpose> <instructions> Use the given context to craft a detailed scenario </instructions> 3. Create Character Persona: <purpose> Create a concise character persona based on provided details. </purpose> <instructions> - Efficiency: Ensure text is machine-readable. - Redaction: Remove unnecessary words. - Format: List. - Context: Maintain conciseness. - Redundancy: Omit redundant details. </instructions> <output> 1. Streamlined, low-context model instructions that are: - Clear and unambiguous - Efficiently structured - Focused on core functionality - Easy to implement and follow. 2. A new scenario with relevant context, details, and list of tags. 3. Concise character persona with essential attributes, remove empty fields. </output>

<purpose> Create new model instructions. </purpose> <instructions> - Context: Utilize user context and goals. - Craft: Clear, concise instructions. - Preserve: Crucial details. - Creative: Offer guidance on roles, rules, response strategies upon request. - Length: Succinct, no fluff. </instructions> <output> New model instructions optimized for clarity and effectiveness. </output>

<purpose> Retrieve lore book information for the user. </purpose> <instructions> - Use only lore book information. - Provide accurate answers without adding, editing, or omitting any information. - Maintain conciseness in responses. </instructions> <output> Verbatim lore book entry value. </output>

<purpose> Condense context for language model readability. </purpose> <instructions> - Ensure efficient and accurate reading. - Ignore human-readable narratives. - Be efficient with wording, concise. - Maintain relevant context. </instructions> <output> Condensed output. </output>

<purpose> Create commands for language model interpretation. </purpose> <instructions> - Use key-value pairs. - Disregard human readability. - Be efficient with wording, concise. - Focus on efficiency, effectiveness, and accuracy for model processing. </instructions> <output> Key: [keyword(s)] <purpose> [Purpose of command] </purpose> <instructions> [Command instructions] </instructions> <output> [Command format] </output> </output>

<purpose> Create keywords from Lorebook entries. </purpose> <instructions> Process the text to identify key terms and phrases that relate to the contents within the lorebook. - Keyword(s) should be contextually relevant and commonly used in natural language. Separate keywords with a comma followed by a space. </instructions> <output> *Keyword(s)* </output>

<purpose> Reformat provided Lorebook entries or context into a custom Markdown (Version 2) </purpose> <instructions> 1. User provides context or Lorebook key. 2. Extract the purpose, instructions, and output from the provided Lorebook entry (if present). 3. If not present, prompt the user for the purpose, instructions, and output. 4. Remove unnecessary tokens. 5. Replace the placeholders in the following template with the extracted/provided information: ``` <purpose> ${purpose} </purpose> <instructions> ${instructions} </instructions> <output> ${output} </output> ``` </instructions> <output> Reformatted Lorebook entry </output>

<purpose> Create lore book entries adhering to formatting guidelines. </purpose> <instructions> - Use key-value pairs. - Disregard human readability. - Be efficient with wording, concise - Focus on efficiency, effectiveness, and accuracy for model processing. - Keyword(s) should be contextually relevant and commonly used in natural language. Separate keywords with a comma followed by a space. </instructions> <output> *Keyword(s)* Value(s). (Nested Value[s]). </output>

<purpose> Create a trigger with an associated effect. </purpose> <instructions> User provides the Purpose, Trigger, and Effect. - If no User context is provided, fabricate. - Conciseness for model readability. - Keyword(s) should be contextually relevant and commonly used in natural language. Separate keywords with a comma followed by a space. - Follow output accurately: </instructions> <output> *Keyword(s)* <purpose> [Purpose of trigger] </purpose> <condition> [Trigger conditions] </condition> <effect> [Effect of trigger] </effect> </output>

<purpose> Merge lorebook entries. </purpose> <instructions> Key-Value Merge Rules: - Update existing keys. - Add new keys. - Merge nested structures. - Keyword(s) should be contextually relevant and commonly used in natural language. Separate keywords with a comma followed by a space. </instructions> <output> *Keyword(s)* Cohesive, merged lorebook entry. </output>

<purpose> Split a provided lorebook entry into the specified number of entries. </purpose> <instructions> 1. Parse the lorebook entry and identify unique keywords. 2. Break the entry into segments based on each unique keyword. 3. Ensure each segment includes any relevant data related to the keyword, such as nested values. - Keyword(s) should be contextually relevant and commonly used in natural language. Separate keywords with a comma followed by a space. 4. Output the new lorebook entries, numbered according to the user's desired split. </instructions> <output> *Keyword(s)*: [Lorebook1] *Keyword(s)*: [Lorebook2] ... *Keyword(s)*: [LorebookN] </output>

<purpose> Create lore book entry using context, then merge it with existing entry. </purpose> 1. Create Lorebook: <instructions> - Use key-value pairs. - Disregard human readability. - Be efficient with wording, concise - Focus on efficiency, effectiveness, and accuracy for model processing. </instructions> 2. Merge Lorebook: <instructions> Key-Value Merge Rules: - Update existing keys. - Add new keys. - Merge nested structures. </instructions> 3. Keyword(s) and output: <instructions> Keyword(s) should be contextually relevant and commonly used in natural language. Separate keywords with a comma followed by a space. </instructions> <output>*Keyword(s)*: Merged lore book entry. </output>

<purpose> Create/optimize scenarios with User context. </purpose> <instructions> Prompt user for necessary details. - Plot: [?]. - Circumstances: [?]. - Natural flow: [?]. - Tags: [?]. </instructions> <output> Create/update scenario with relevant details and list of tags. </output>

<purpose> Create or refine character personas. </purpose> <instructions> - Ensure efficient, accurate reading. - Ignore human readability. - Be efficient with wording, concise. - Maintain relevant context. - Omit redundancy. - Follow the format of CharTemplate (Fabricate empty fields): </instructions> <output> *Name* [Age], [Gender] [Occupation], [Schedule] Physical Appearance: - [Height] - [Hair] - [Beard] - [Eyes] - [Build] - [Demeanor] Speech and Communication: - [Tone] - [Speech Patterns] - [Emotional Expression] Relationships: - Friends: [Friend 1], [Friend 2] Family: - Family: [Family 1], [Family 2], [Family 3] Likes: - [Interest 1], [Interest 2], [Interest 3], [Interest 4], [Interest 5] Dislikes: - [Dislike 1], [Dislike 2], [Dislike 3] Psychology: - [Complexity/Emotion 1], [Complexity/Emotion 2], [Complexity/Emotion 3], [Complexity/Emotion 4] Background: - [Background 1], [Background 2] Secrets: - [Secret 1], [Secret 2] Goals: - [Goal 1], [Goal 2] </output>

<purpose> Create character entries from excerpt. </purpose> <instructions> Context: User provided excerpt. Focus: - Traits: Appearance, speech, mannerisms. - Relations: Character dynamics. - Goals: Motivations. </instructions> <output> Condensed, efficient entries: *Name*: [Name] *Relation*: [Character dynamic] *Speech*: [Style/tone] *Mannerisms*: [Actions/gestures] *Goals*: [Motivations/desires] </output>

<purpose> Create or edit a scene using context provided by the user or fabricate one if none is given. </purpose> <instructions> 1. If user provides context, then extract relevant details. 2. Interpret details to describe characters: movements, appearances, clothing, and actions. 3. Incorporate sensory details (smell, texture, taste, feelings). 4. Create a highly immersive scene with the provided context. 5. If no context is provided, generate a new scene using the model's understanding of common fantasy tropes and themes. </instructions> <output> A highly detailed scene. </output>

<purpose> Generate an introductory message for a character based on user-provided context. <instructions> The "First Message" will follow these guidelines: - Creative Approach: Setup through character interaction or monologue. - Natural integration of scenario, location, and relationships. - Demonstrate response style, format, and content. - Create an immersive excerpt in novel format. </instructions> <output> Response Format: Narration "Dialogue" says Character non-verbal "Dialogue". </output>

<purpose> Generate dialogue based on user-provided character context; If no context is given, fabricate a character. </purpose> <instructions> - Focus on dialogue. - Context: Portray speech, mannerisms accurately. - Age: Reflect age in dialogue/actions. - Ambiguity: Avoid plot points, use interests/likes. - Description: Include recurring description. - Format: Quotes for dialogue, no indicators for non-verbal. - Create immersive dialogue using only the character specified or created. </instructions> <output> Response format: `#{Character}:` "Dialogue" says Character non-verbal "Dialogue". </output>

<purpose>Build a world context for fictional settings</purpose> <instructions>List of world-building elements to include:</instructions> <output> `Background: [element1] [element2] ... [elementN]` - Sets world background `Lore: [element1] [element2] ... [elementN]` - Sets world lore `Rules: [element1] [element2] ... [elementN]` - Sets world rules and restrictions `Geography: [element1] [element2] ... [elementN]` - Sets world geography `Population: [element1] [element2] ... [elementN]` - Sets world population and inhabitants `Technology: [element1] [element2] ... [elementN]` - Sets world technology level `Magic: [element1] [element2] ... [elementN]` - Sets world magic system `Flora: [element1] [element2] ... [elementN]` - Sets world flora `Fauna: [element1] [element2] ... [elementN]` - Sets world fauna `Climate: [element1] [element2] ... [elementN]` - Sets world climate </output>

<purpose> Create or refine a town. </purpose> <instructions> - Use context provided by the user to create or refine a setting. - If no context is given, fabricate a fictional place. </instructions> <output> *[Type of location]* *Name*: - [Location] - [Population and Culture] - [Government] - [Challenges] - [Points of interest] </output>

<purpose> Create or refine a place of interest. </purpose> <instructions> - Use context provided by the user to create or refine a setting. - If no context is given, fabricate a fictional place. </instructions> <output> *TypeOfPlace* *Name*: - Description - Summary </output>

<purpose> Create or refine an event or gathering. </purpose> <instructions> - Use context provided by the user to create or refine an event. - If no context is given, fabricate details. </instructions> <output> *Location* *Name*: - Description - Activities - Significance - Summary </output>

<purpose> Create a new faction within the high medieval fantasy setting </purpose> <instructions> Provide the following information; if no context is provided, fabricate: 1. Faction name 2. Faction description 3. Faction emblem (if available) 4. Faction color (if available) 5. Faction leader (name and description) 6. Faction goals 7. Faction members (initial list of characters) 8. Faction location (city, region, or kingdom) 9. Faction type (mercenaries, knights, thieves, clerics, etc.) 10. Any additional relevant information (alliances, rivals, etc.) </instructions> <output> Created faction </output>

<purpose>Guide a low-context LLM in generating comprehensive, visually-oriented AI image prompts</purpose> <instructions> Create detailed, visually-focused image prompts following these rules: 1. Use 15-25 comma-separated words or short phrases 2. Include these visual elements: - Subject's sex (male, female, androgynous) - Appearance details (hair, eyes, build) - Attire and accessories - Pose and expression - Setting and environment - Camera perspective (e.g., close-up, full body, bird's eye view) - Lighting - Art style 3. Omit non-visual information (e.g., personality traits, background story) 4. Prioritize specific, descriptive visual details 5. Include one unique visual element for distinctiveness 6. Use this structure: sex, subject details, attire, pose, setting, camera perspective, style, lighting Avoid mentioning: - Character names - Psychological traits - Non-visual goals or motivations - Family information - Specific age (unless visually apparent) </instructions> <output> `Prompt: [prompt]` Where `[prompt]` is the generated image prompt focusing solely on visual elements, including sex and camera perspective. ``` Example: `Prompt: male, young adult, long flowing white hair, piercing grey eyes, toned slim build, medieval hunting attire, bow and quiver, standing alertly, lush dense forest, dappled sunlight, three-quarter view, anime style, dramatic side lighting` ``` </output>

<purpose> Elaborate on user-specified topics. </purpose> <instructions> - Provide concise, accurate responses. - Ensure human readability. </instructions> <output> Answer to user's query, elaborating on the specified topic. </output>

<purpose> Extract tone, theme, mood, atmosphere, and style from the provided context. </purpose> <instructions> Parse the given excerpt for keywords and phrases indicative of tone, theme, mood, atmosphere, and style. </instructions> <output> [Output] </output>

<purpose> This is meant as a reference for model priorities and context rules. </purpose> <output> Priority & Category: Permanent Context, Forgettable Context, Variable Context. Permanent Context: - *Model Instructions*: Define style, format, and language for consistent responses. Prime with document types: transcripts, scripts, RP transcripts. - *Model Behavior*: Interpret and follow instructions precisely. Adapt responses to the specified style, format, and language. - *Character Persona*: Include character details such as motivations, flaws, and relationships. - *User Persona*: Understand user goals and preferences, including vivid descriptions, narrative pushes, POV choices, 'beg and bribe' techniques, and OOC notes. - *Scenario*: Use positive framing and consider narrative pacing. Provide a background and starting point for the interaction, shared goal, or activity. Avoid time-specific details and location specifics in permanent context. Forgettable Context: - *First Message* and *Old Chat History*: Forget to ensure a fresh start and prevent conflicts. Variable Context: - *Chat History*, *Lore Entries*, *Author's Notes*: Vital dynamic context sources. - *Example Dialogue*: Impacts form/style of responses. Provide at least three examples, over-emphasizing desired qualities. - *Reinforcement*: Reinforce out-of-context instructions with multiple examples to emphasize desired qualities and ensure model comprehension. </output>

<purpose> Emulate time passage. </purpose> <condition> {user} sleeps. </condition> <instructions> Advance day. </instructions> <output> Day: [X], Weekday: [Y]. </output> #EndResponse